/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 1346 by ph10, Tue Jul 2 18:37:36 2013 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.34 xx-xxxx-201x
5    --------------------------
6    
7    1.  Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
8        executable memory. Patch inspired by Carsten Klein.
9    
10    2.  ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
11        this macro is never tested and has no effect, because the work to support
12        coverage involves only compiling and linking options and special targets in
13        the Makefile. The comment in config.h implied that defining the macro would
14        enable coverage support, which is totally false. There was also support for
15        setting this macro in the CMake files (my fault, I just copied it from
16        configure). SUPPORT_GCOV has now been removed.
17    
18    3.  Make a small performance improvement in strlen16() and strlen32() in
19        pcretest.
20    
21    4.  Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
22        detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
23        these cases). There was also one in pcretest.c.
24    
25    5.  Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
26    
27    
28    Version 8.33 28-May-2013
29    --------------------------
30    
31    1.  Added 'U' to some constants that are compared to unsigned integers, to
32        avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
33        variables that are added to signed variables, to ensure the result is
34        signed and can be negated.
35    
36    2.  Applied patch by Daniel Richard G for quashing MSVC warnings to the
37        CMake config files.
38    
39    3.  Revise the creation of config.h.generic so that all boolean macros are
40        #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
41        overriding via -D on the command line possible.
42    
43    4.  Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
44        to unsigned int is reported to make a quite noticeable speed difference in
45        a specific Windows environment. Testing on Linux did also appear to show
46        some benefit (and it is clearly not harmful). Also fixed the definition of
47        Xop which should be unsigned.
48    
49    5.  Related to (4), changing the definition of the intermediate variable cc
50        in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
51        improvements.
52    
53    6.  Fix forward search in JIT when link size is 3 or greater. Also removed some
54        unnecessary spaces.
55    
56    7.  Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
57        and later.
58    
59    8.  Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
60    
61    9.  Optimizing fast_forward_start_bits in JIT.
62    
63    10. Adding support for callouts in JIT, and fixing some issues revealed
64        during this work. Namely:
65    
66        (a) Unoptimized capturing brackets incorrectly reset on backtrack.
67    
68        (b) Minimum length was not checked before the matching is started.
69    
70    11. The value of capture_last that is passed to callouts was incorrect in some
71        cases when there was a capture on one path that was subsequently abandoned
72        after a backtrack. Also, the capture_last value is now reset after a
73        recursion, since all captures are also reset in this case.
74    
75    12. The interpreter no longer returns the "too many substrings" error in the
76        case when an overflowing capture is in a branch that is subsequently
77        abandoned after a backtrack.
78    
79    13. In the pathological case when an offset vector of size 2 is used, pcretest
80        now prints out the matched string after a yield of 0 or 1.
81    
82    14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
83        Only supported by the JIT compiler at the moment.
84    
85    15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
86    
87    16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
88        offset of the starting point of the matching process, provided the offsets
89        vector is large enough.
90    
91    17. The \A escape now records a lookbehind value of 1, though its execution
92        does not actually inspect the previous character. This is to ensure that,
93        in partial multi-segment matching, at least one character from the old
94        segment is retained when a new segment is processed. Otherwise, if there
95        are no lookbehinds in the pattern, \A might match incorrectly at the start
96        of a new segment.
97    
98    18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
99    
100    19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
101        gives some modest performance improvement in 8-bit mode.
102    
103    20. Added the PCRE-specific property \p{Xuc} for matching characters that can
104        be expressed in certain programming languages using Universal Character
105        Names.
106    
107    21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
108        which points out that "non characters" are not "characters that may not
109        appear in Unicode strings" but rather "characters that are reserved for
110        internal use and have only local meaning".
111    
112    22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
113        there was a conditional group that depended on an assertion, if the
114        assertion was false, the callout that immediately followed the alternation
115        in the condition was skipped when pcre_exec() was used for matching.
116    
117    23. Allow an explicit callout to be inserted before an assertion that is the
118        condition for a conditional group, for compatibility with automatic
119        callouts, which always insert a callout at this point.
120    
121    24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
122        confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
123    
124    25. (*PRUNE) is now supported by the JIT compiler.
125    
126    26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
127    
128    27. Fix the case where there are two or more SKIPs with arguments that may be
129        ignored.
130    
131    28. (*SKIP) is now supported by the JIT compiler.
132    
133    29. (*THEN) is now supported by the JIT compiler.
134    
135    30. Update RunTest with additional test selector options.
136    
137    31. The way PCRE handles backtracking verbs has been changed in two ways.
138    
139        (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
140        SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
141        backtracking. In some cases this makes it more Perl-compatible, but Perl's
142        rather obscure rules do not always do the same thing.
143    
144        (2) Previously, backtracking verbs were confined within assertions. This is
145        no longer the case for positive assertions, except for (*ACCEPT). Again,
146        this sometimes improves Perl compatibility, and sometimes does not.
147    
148    32. A number of tests that were in test 2 because Perl did things differently
149        have been moved to test 1, because either Perl or PCRE has changed, and
150        these tests are now compatible.
151    
152    32. Backtracking control verbs are now handled in the same way in JIT and
153        interpreter.
154    
155    33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
156        contained a forward subroutine reference caused a compile error.
157    
158    34. Auto-detect and optimize limited repetitions in JIT.
159    
160    35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
161        blocking (*UTF) etc.
162    
163    36. In the interpreter, maximizing pattern repetitions for characters and
164        character types now use tail recursion, which reduces stack usage.
165    
166    37. The value of the max lookbehind was not correctly preserved if a compiled
167        and saved regex was reloaded on a host of different endianness.
168    
169    38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
170        of the compiled pattern block, expand the flags field from 16 to 32 bits
171        because it was almost full.
172    
173    39. Try madvise first before posix_madvise.
174    
175    40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
176        with a pattern such as ^$. It has taken 4 years for anybody to notice! The
177        original change locked out all matches of empty strings. This has been
178        changed so that one match of an empty string per line is recognized.
179        Subsequent searches on the same line (for colouring or for --only-matching,
180        for example) do not recognize empty strings.
181    
182    41. Applied a user patch to fix a number of spelling mistakes in comments.
183    
184    42. Data lines longer than 65536 caused pcretest to crash.
185    
186    43. Clarified the data type for length and startoffset arguments for pcre_exec
187        and pcre_dfa_exec in the function-specific man pages, where they were
188        explicitly stated to be in bytes, never having been updated. I also added
189        some clarification to the pcreapi man page.
190    
191    44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
192        a segmentation fault.
193    
194    
195    Version 8.32 30-November-2012
196    -----------------------------
197    
198    1.  Improved JIT compiler optimizations for first character search and single
199        character iterators.
200    
201    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
202        Patch by Daniel Richard G.
203    
204    3.  Single character iterator optimizations in the JIT compiler.
205    
206    4.  Improved JIT compiler optimizations for character ranges.
207    
208    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
209        Reported by Giuseppe D'Angelo.
210    
211    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
212        of a line, was being set incorrectly in cases where .* appeared inside
213        atomic brackets at the start of a pattern, or where there was a subsequent
214        *PRUNE or *SKIP.
215    
216    7.  Improved instruction cache flush for POWER/PowerPC.
217        Patch by Daniel Richard G.
218    
219    8.  Fixed a number of issues in pcregrep, making it more compatible with GNU
220        grep:
221    
222        (a) There is now no limit to the number of patterns to be matched.
223    
224        (b) An error is given if a pattern is too long.
225    
226        (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
227            are now supported.
228    
229        (d) --exclude-from and --include-from (multiple use) have been added.
230    
231        (e) Exclusions and inclusions now apply to all files and directories, not
232            just to those obtained from scanning a directory recursively.
233    
234        (f) Multiple uses of -f and --file-list are now supported.
235    
236        (g) In a Windows environment, the default for -d has been changed from
237            "read" (the GNU grep default) to "skip", because otherwise the presence
238            of a directory in the file list provokes an error.
239    
240        (h) The documentation has been revised and clarified in places.
241    
242    9.  Improve the matching speed of capturing brackets.
243    
244    10. Changed the meaning of \X so that it now matches a Unicode extended
245        grapheme cluster.
246    
247    11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
248        out POSIX threads when JIT support is configured.
249    
250    12. Added support for PCRE_STUDY_EXTRA_NEEDED.
251    
252    13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
253        structure could go wrong in environments where size_t is not the same size
254        as int.
255    
256    14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
257        set.
258    
259    15. The EBCDIC support had decayed; later updates to the code had included
260        explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
261        general tidy up of EBCDIC-related issues, and the documentation was also
262        not quite right. There is now a test that can be run on ASCII systems to
263        check some of the EBCDIC-related things (but is it not a full test).
264    
265    16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
266        in a small tidy to the code.
267    
268    17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
269    
270    18. If the --only-matching (-o) option in pcregrep is specified multiple
271        times, each one causes appropriate output. For example, -o1 -o2 outputs the
272        substrings matched by the 1st and 2nd capturing parentheses. A separating
273        string can be specified by --om-separator (default empty).
274    
275    19. Improving the first n character searches.
276    
277    20. Turn case lists for horizontal and vertical white space into macros so that
278        they are defined only once.
279    
280    21. This set of changes together give more compatible Unicode case-folding
281        behaviour for characters that have more than one other case when UCP
282        support is available.
283    
284        (a) The Unicode property table now has offsets into a new table of sets of
285            three or more characters that are case-equivalent. The MultiStage2.py
286            script that generates these tables (the pcre_ucd.c file) now scans
287            CaseFolding.txt instead of UnicodeData.txt for character case
288            information.
289    
290        (b) The code for adding characters or ranges of characters to a character
291            class has been abstracted into a generalized function that also handles
292            case-independence. In UTF-mode with UCP support, this uses the new data
293            to handle characters with more than one other case.
294    
295        (c) A bug that is fixed as a result of (b) is that codepoints less than 256
296            whose other case is greater than 256 are now correctly matched
297            caselessly. Previously, the high codepoint matched the low one, but not
298            vice versa.
299    
300        (d) The processing of \h, \H, \v, and \ in character classes now makes use
301            of the new class addition function, using character lists defined as
302            macros alongside the case definitions of 20 above.
303    
304        (e) Caseless back references now work with characters that have more than
305            one other case.
306    
307        (f) General caseless matching of characters with more than one other case
308            is supported.
309    
310    22. Unicode character properties were updated from Unicode 6.2.0
311    
312    23. Improved CMake support under Windows. Patch by Daniel Richard G.
313    
314    24. Add support for 32-bit character strings, and UTF-32
315    
316    25. Major JIT compiler update (code refactoring and bugfixing).
317        Experimental Sparc 32 support is added.
318    
319    26. Applied a modified version of Daniel Richard G's patch to create
320        pcre.h.generic and config.h.generic by "make" instead of in the
321        PrepareRelease script.
322    
323    27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
324        pcre_compile.c when checking for a zero character.
325    
326    28. Introducing a native interface for JIT. Through this interface, the compiled
327        machine code can be directly executed. The purpose of this interface is to
328        provide fast pattern matching, so several sanity checks are not performed.
329        However, feature tests are still performed. The new interface provides
330        1.4x speedup compared to the old one.
331    
332    29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
333        the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
334        was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
335    
336    30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
337        greater than 0x7fffffff (which cannot be represented in UTF-8, even under
338        the "old" RFC 2279). Instead, it ended up passing a negative length to
339        pcre_exec().
340    
341    31. Add support for GCC's visibility feature to hide internal functions.
342    
343    32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
344        "unknown -C option" after outputting 0 or 1.
345    
346    33. There is now support for generating a code coverage report for the test
347        suite in environments where gcc is the compiler and lcov is installed. This
348        is mainly for the benefit of the developers.
349    
350    34. If PCRE is built with --enable-valgrind, certain memory regions are marked
351        unaddressable using valgrind annotations, allowing valgrind to detect
352        invalid memory accesses. This is mainly for the benefit of the developers.
353    
354    25. (*UTF) can now be used to start a pattern in any of the three libraries.
355    
356    26. Give configure error if --enable-cpp but no C++ compiler found.
357    
358    
359    Version 8.31 06-July-2012
360    -------------------------
361    
362    1.  Fixing a wrong JIT test case and some compiler warnings.
363    
364    2.  Removed a bashism from the RunTest script.
365    
366    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
367        to unsigned type, result still unsigned" that was given by an MS compiler
368        on encountering the code "-sizeof(xxx)".
369    
370    4.  Partial matching support is added to the JIT compiler.
371    
372    5.  Fixed several bugs concerned with partial matching of items that consist
373        of more than one character:
374    
375        (a) /^(..)\1/ did not partially match "aba" because checking references was
376            done on an "all or nothing" basis. This also applied to repeated
377            references.
378    
379        (b) \R did not give a hard partial match if \r was found at the end of the
380            subject.
381    
382        (c) \X did not give a hard partial match after matching one or more
383            characters at the end of the subject.
384    
385        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
386            a partial match for the string "\r".
387    
388        (e) When newline was set to CRLF, the metacharacter "." did not recognize
389            a partial match for a CR character at the end of the subject string.
390    
391    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
392        running pcretest, the text "(JIT)" added to the output whenever JIT is
393        actually used to run the match.
394    
395    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
396        or /S+[+] with a digit between 1 and 7.
397    
398    8.  OP_NOT now supports any UTF character not just single-byte ones.
399    
400    9.  (*MARK) control verb is now supported by the JIT compiler.
401    
402    10. The command "./RunTest list" lists the available tests without actually
403        running any of them. (Because I keep forgetting what they all are.)
404    
405    11. Add PCRE_INFO_MAXLOOKBEHIND.
406    
407    12. Applied a (slightly modified) user-supplied patch that improves performance
408        when the heap is used for recursion (compiled with --disable-stack-for-
409        recursion). Instead of malloc and free for each heap frame each time a
410        logical recursion happens, frames are retained on a chain and re-used where
411        possible. This sometimes gives as much as 30% improvement.
412    
413    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
414        call.
415    
416    14. As documented, (*COMMIT) is now confined to within a positive assertion.
417    
418    15. It is now possible to link pcretest with libedit as an alternative to
419        libreadline.
420    
421    16. (*COMMIT) control verb is now supported by the JIT compiler.
422    
423    17. The Unicode data tables have been updated to Unicode 6.1.0.
424    
425    18. Added --file-list option to pcregrep.
426    
427    19. Added binary file support to pcregrep, including the -a, --binary-files,
428        -I, and --text options.
429    
430    20. The madvise function is renamed for posix_madvise for QNX compatibility
431        reasons. Fixed by Giuseppe D'Angelo.
432    
433    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
434    
435    22. Changed the option for creating symbolic links for 16-bit man pages from
436        -s to -sf so that re-installing does not cause issues.
437    
438    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
439    
440    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
441        matches in certain environments (the workspace was not being correctly
442        retained). Also added to pcre_dfa_exec() a simple plausibility check on
443        some of the workspace data at the beginning of a restart.
444    
445    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
446        was not doing so when it should - probably a typo introduced by SVN 528
447        (change 8.10/14).
448    
449    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
450        \w+ when the character tables indicated that \x{c4} was a word character.
451        There were several related cases, all because the tests for doing a table
452        lookup were testing for characters less than 127 instead of 255.
453    
454    27. If a pattern contains capturing parentheses that are not used in a match,
455        their slots in the ovector are set to -1. For those that are higher than
456        any matched groups, this happens at the end of processing. In the case when
457        there were back references that the ovector was too small to contain
458        (causing temporary malloc'd memory to be used during matching), and the
459        highest capturing number was not used, memory off the end of the ovector
460        was incorrectly being set to -1. (It was using the size of the temporary
461        memory instead of the true size.)
462    
463    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
464        ovector size, it uses memory at the end of the block that it has got.
465    
466    29. Check for an overlong MARK name and give an error at compile time. The
467        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
468    
469    30. JIT compiler update.
470    
471    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
472        Rill for the patch.
473    
474    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
475    
476    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
477    
478    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
479        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
480        was enabled.
481    
482    35. Improve JIT code generation for greedy plus quantifier.
483    
484    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
485        1 to "aa" instead of to an empty string. The bug affected repeated groups
486        that could potentially match an empty string.
487    
488    37. Optimizing single character iterators in JIT.
489    
490    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
491        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
492        codepoints that are too big for the mode are faulted, and in a UTF mode,
493        disallowed codepoints are also faulted.
494    
495    39. If PCRE was compiled with UTF support, in three places in the DFA
496        matcher there was code that should only have been obeyed in UTF mode, but
497        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
498        processing when bytes with values greater than 127 were present. In 16-bit
499        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
500        both cases the values are those that cannot be the first data item in a UTF
501        character. The three items that might have provoked this were recursions,
502        possessively repeated groups, and atomic groups.
503    
504    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
505        and pcregrep, because some OS require shared objects to be explicitly
506        passed to ld, causing the link step to fail if they are not.
507    
508    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
509        mode, patterns that started with \h* or \R* might be incorrectly matched.
510    
511    
512    Version 8.30 04-February-2012
513    -----------------------------
514    
515    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
516        name is defined in ctype.h.
517    
518    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
519        only in quite long subpatterns.
520    
521    3.  Removed the function pcre_info(), which has been obsolete and deprecated
522        since it was replaced by pcre_fullinfo() in February 2000.
523    
524    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
525        match a (*MARK), and the match failed at the start of the subject, a
526        reference to memory before the start of the subject could occur. This bug
527        was introduced by fix 17 of release 8.21.
528    
529    5.  A reference to an unset group with zero minimum repetition was giving
530        totally wrong answers (in non-JavaScript-compatibility mode). For example,
531        /(another)?(\1?)test/ matched against "hello world test". This bug was
532        introduced in release 8.13.
533    
534    6.  Add support for 16-bit character strings (a large amount of work involving
535        many changes and refactorings).
536    
537    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
538        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
539        from a file.
540    
541    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
542        rounding is not applied in this particular case).
543    
544    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
545        if they appear, or are escaped, in patterns.
546    
547    10. Get rid of a number of -Wunused-but-set-variable warnings.
548    
549    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
550        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
551        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
552        also returns the mark "x". This bug applied to capturing parentheses,
553        non-capturing parentheses, and atomic parentheses. It also applied to some
554        assertions.
555    
556    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
557        information out of configure.ac instead of relying on pcre.h.generic, which
558        is not stored in the repository.
559    
560    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
561        -lreadline.
562    
563    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
564    
565    15. Applied Graycode's patch to put the top-level frame on the stack rather
566        than the heap when not using the stack for recursion. This gives a
567        performance improvement in many cases when recursion is not deep.
568    
569    16. Experimental code added to "pcretest -C" to output the stack frame size.
570    
571    
572    Version 8.21 12-Dec-2011
573    ------------------------
574    
575    1.  Updating the JIT compiler.
576    
577    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
578        are added as well.
579    
580    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
581        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
582        calling _pcre_jit_exec. Some extra comments are added.
583    
584    4.  (*MARK) settings inside atomic groups that do not contain any capturing
585        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
586        was introduced by change 18 for 8.20.
587    
588    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
589        ECMA-262 standard.
590    
591    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
592        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
593        This bug was probably introduced by change 9 of 8.13.
594    
595    7.  While fixing 6 above, I noticed that a number of other items were being
596        incorrectly rejected as "not fixed length". This arose partly because newer
597        opcodes had not been added to the fixed-length checking code. I have (a)
598        corrected the bug and added tests for these items, and (b) arranged for an
599        error to occur if an unknown opcode is encountered while checking for fixed
600        length instead of just assuming "not fixed length". The items that were
601        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
602        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
603        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
604    
605    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
606        being incorrectly compiled and would have given unpredicatble results.
607    
608    9.  A possessively repeated subpattern with minimum repeat count greater than
609        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
610        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
611        the first (A) could occur when it should not.
612    
613    10. Add a cast and remove a redundant test from the code.
614    
615    11. JIT should use pcre_malloc/pcre_free for allocation.
616    
617    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
618        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
619        is anything other than /usr, -L is still shown).
620    
621    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
622    
623    14. Perl does not support \N without a following name in a [] class; PCRE now
624        also gives an error.
625    
626    15. If a forward reference was repeated with an upper limit of around 2000,
627        it caused the error "internal error: overran compiling workspace". The
628        maximum number of forward references (including repeats) was limited by the
629        internal workspace, and dependent on the LINK_SIZE. The code has been
630        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
631        the default depends on LINK_SIZE. There is a new upper limit (for safety)
632        of around 200,000 forward references. While doing this, I also speeded up
633        the filling in of repeated forward references.
634    
635    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
636        incorrectly expecting the subject to contain another "a" after the start.
637    
638    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
639        in the match, the SKIP should be ignored. This was not happening; instead
640        the SKIP was being treated as NOMATCH. For patterns such as
641        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
642        tested.
643    
644    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
645        now much more compatible with Perl, in particular in cases where the result
646        is a non-match for a non-anchored pattern. For example, if
647        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
648        "m", where previously it did not return a name. A side effect of this
649        change is that for partial matches, the last encountered mark name is
650        returned, as for non matches. A number of tests that were previously not
651        Perl-compatible have been moved into the Perl-compatible test files. The
652        refactoring has had the pleasing side effect of removing one argument from
653        the match() function, thus reducing its stack requirements.
654    
655    19. If the /S+ option was used in pcretest to study a pattern using JIT,
656        subsequent uses of /S (without +) incorrectly behaved like /S+.
657    
658    21. Retrieve executable code size support for the JIT compiler and fixing
659        some warnings.
660    
661    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
662        not work when the shorter character appeared right at the end of the
663        subject string.
664    
665    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
666        systems.
667    
668    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
669        output it when the /M option is used in pcretest.
670    
671    25. The CheckMan script was not being included in the distribution. Also, added
672        an explicit "perl" to run Perl scripts from the PrepareRelease script
673        because this is reportedly needed in Windows.
674    
675    26. If study data was being save in a file and studying had not found a set of
676        "starts with" bytes for the pattern, the data written to the file (though
677        never used) was taken from uninitialized memory and so caused valgrind to
678        complain.
679    
680    27. Updated RunTest.bat as provided by Sheri Pierce.
681    
682    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
683    
684    29. Computation of memory usage for the table of capturing group names was
685        giving an unnecessarily large value.
686    
687    
688    Version 8.20 21-Oct-2011
689    ------------------------
690    
691    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
692        a POSIX class. After further experiments with Perl, which convinced me that
693        Perl has bugs and confusions, a closing square bracket is no longer allowed
694        in a POSIX name. This bug also affected patterns with classes that started
695        with full stops.
696    
697    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
698        captured substring, but while checking the failing first alternative,
699        substring 1 is temporarily captured. If the output vector supplied to
700        pcre_exec() was not big enough for this capture, the yield of the function
701        was still zero ("insufficient space for captured substrings"). This cannot
702        be totally fixed without adding another stack variable, which seems a lot
703        of expense for a edge case. However, I have improved the situation in cases
704        such as /(a)(b)x|abc/ matched against "abc", where the return code
705        indicates that fewer than the maximum number of slots in the ovector have
706        been set.
707    
708    3.  Related to (2) above: when there are more back references in a pattern than
709        slots in the output vector, pcre_exec() uses temporary memory during
710        matching, and copies in the captures as far as possible afterwards. It was
711        using the entire output vector, but this conflicts with the specification
712        that only 2/3 is used for passing back captured substrings. Now it uses
713        only the first 2/3, for compatibility. This is, of course, another edge
714        case.
715    
716    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
717        main code base, and can be used by building with --enable-jit. When this is
718        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
719        runtime --no-jit option is given.
720    
721    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
722        ovector, the return from the function was zero, implying that there were
723        other matches that did not fit. The correct "exactly full" value is now
724        returned.
725    
726    6.  If a subpattern that was called recursively or as a subroutine contained
727        (*PRUNE) or any other control that caused it to give a non-standard return,
728        invalid errors such as "Error -26 (nested recursion at the same subject
729        position)" or even infinite loops could occur.
730    
731    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
732        computing the minimum length on reaching *ACCEPT, and so ended up with the
733        wrong value of 1 rather than 0. Further investigation indicates that
734        computing a minimum subject length in the presence of *ACCEPT is difficult
735        (think back references, subroutine calls), and so I have changed the code
736        so that no minimum is registered for a pattern that contains *ACCEPT.
737    
738    8.  If (*THEN) was present in the first (true) branch of a conditional group,
739        it was not handled as intended. [But see 16 below.]
740    
741    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
742        Sheri Pierce.
743    
744    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
745        the first byte in a match must be "a".
746    
747    11. Change 17 for 8.13 increased the recursion depth for patterns like
748        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
749        pattern contains any instances of (*THEN). If it does not, the old
750        optimizations are restored. It would be nice to do this on a per-group
751        basis, but at the moment that is not feasible.
752    
753    12. In some environments, the output of pcretest -C is CRLF terminated. This
754        broke RunTest's code that checks for the link size. A single white space
755        character after the value is now allowed for.
756    
757    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
758        For "fr", it uses the Windows-specific input and output files.
759    
760    14. If (*THEN) appeared in a group that was called recursively or as a
761        subroutine, it did not work as intended. [But see next item.]
762    
763    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
764        pattern fragments (but not containing any | characters). If A and B are
765        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
766        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
767        D. In other words, Perl considers parentheses that do not contain any |
768        characters to be part of a surrounding alternative, whereas PCRE was
769        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
770        differently. PCRE now behaves in the same way as Perl, except in the case
771        of subroutine/recursion calls such as (?1) which have in any case always
772        been different (but PCRE had them first :-).
773    
774    16. Related to 15 above: Perl does not treat the | in a conditional group as
775        creating alternatives. Such a group is treated in the same way as an
776        ordinary group without any | characters when processing (*THEN). PCRE has
777        been changed to match Perl's behaviour.
778    
779    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
780        RunGrepTest script failed.
781    
782    18. Change 22 for version 13 caused atomic groups to use more stack. This is
783        inevitable for groups that contain captures, but it can lead to a lot of
784        stack use in large patterns. The old behaviour has been restored for atomic
785        groups that do not contain any capturing parentheses.
786    
787    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
788        suppress the check for a minimum subject length at run time. (If it was
789        given to pcre_exec() or pcre_dfa_exec() it did work.)
790    
791    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
792        fail to work when decoding hex characters in data strings in EBCDIC
793        environments.
794    
795    21. It appears that in at least one Mac OS environment, the isxdigit() function
796        is implemented as a macro that evaluates to its argument more than once,
797        contravening the C 90 Standard (I haven't checked a later standard). There
798        was an instance in pcretest which caused it to go wrong when processing
799        \x{...} escapes in subject strings. The has been rewritten to avoid using
800        things like p++ in the argument of isxdigit().
801    
802    
803    Version 8.13 16-Aug-2011
804    ------------------------
805    
806    1.  The Unicode data tables have been updated to Unicode 6.0.0.
807    
808    2.  Two minor typos in pcre_internal.h have been fixed.
809    
810    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
811        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
812        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
813    
814    4.  There were a number of related bugs in the code for matching backrefences
815        caselessly in UTF-8 mode when codes for the characters concerned were
816        different numbers of bytes. For example, U+023A and U+2C65 are an upper
817        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
818        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
819        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
820        2-byte code at the end of the subject (it thought there wasn't enough data
821        left).
822    
823    5.  Comprehensive information about what went wrong is now returned by
824        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
825        as the output vector has at least 2 elements. The offset of the start of
826        the failing character and a reason code are placed in the vector.
827    
828    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
829        now returned is for the first byte of the failing character, instead of the
830        last byte inspected. This is an incompatible change, but I hope it is small
831        enough not to be a problem. It makes the returned offset consistent with
832        pcre_exec() and pcre_dfa_exec().
833    
834    7.  pcretest now gives a text phrase as well as the error number when
835        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
836        failure, the offset and reason code are output.
837    
838    8.  When \R was used with a maximizing quantifier it failed to skip backwards
839        over a \r\n pair if the subsequent match failed. Instead, it just skipped
840        back over a single character (\n). This seems wrong (because it treated the
841        two characters as a single entity when going forwards), conflicts with the
842        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
843        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
844        has been changed.
845    
846    9.  Some internal refactoring has changed the processing so that the handling
847        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
848        time (the PCRE_DOTALL option was changed this way some time ago: version
849        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
850        which was always a bit of a fudge. It also means that there is one less
851        argument for the match() function, which reduces its stack requirements
852        slightly. This change also fixes an incompatibility with Perl: the pattern
853        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
854    
855    10. More internal refactoring has drastically reduced the number of recursive
856        calls to match() for possessively repeated groups such as (abc)++ when
857        using pcre_exec().
858    
859    11. While implementing 10, a number of bugs in the handling of groups were
860        discovered and fixed:
861    
862        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
863        (a|)*(?1) gave a compile-time internal error.
864        ((a|)+)+  did not notice that the outer group could match an empty string.
865        (^a|^)+   was not marked as anchored.
866        (.*a|.*)+ was not marked as matching at start or after a newline.
867    
868    12. Yet more internal refactoring has removed another argument from the match()
869        function. Special calls to this function are now indicated by setting a
870        value in a variable in the "match data" data block.
871    
872    13. Be more explicit in pcre_study() instead of relying on "default" for
873        opcodes that mean there is no starting character; this means that when new
874        ones are added and accidentally left out of pcre_study(), testing should
875        pick them up.
876    
877    14. The -s option of pcretest has been documented for ages as being an old
878        synonym of -m (show memory usage). I have changed it to mean "force study
879        for every regex", that is, assume /S for every regex. This is similar to -i
880        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
881        using it. It makes it easier to run collections of tests with and without
882        study enabled, and thereby test pcre_study() more easily. All the standard
883        tests are now run with and without -s (but some patterns can be marked as
884        "never study" - see 20 below).
885    
886    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
887        restoration of the capturing data to the outer values was not happening
888        correctly.
889    
890    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
891        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
892        pattern had matched an empty string, and so incorrectly returned a no
893        match.
894    
895    17. There was optimizing code for the last branch of non-capturing parentheses,
896        and also for the obeyed branch of a conditional subexpression, which used
897        tail recursion to cut down on stack usage. Unfortunately, now that there is
898        the possibility of (*THEN) occurring in these branches, tail recursion is
899        no longer possible because the return has to be checked for (*THEN). These
900        two optimizations have therefore been removed. [But see 8.20/11 above.]
901    
902    18. If a pattern containing \R was studied, it was assumed that \R always
903        matched two bytes, thus causing the minimum subject length to be
904        incorrectly computed because \R can also match just one byte.
905    
906    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
907        was incorrectly computed.
908    
909    20. If /S is present twice on a test pattern in pcretest input, it now
910        *disables* studying, thereby overriding the use of -s on the command line
911        (see 14 above). This is necessary for one or two tests to keep the output
912        identical in both cases.
913    
914    21. When (*ACCEPT) was used in an assertion that matched an empty string and
915        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
916    
917    22. When an atomic group that contained a capturing parenthesis was
918        successfully matched, but the branch in which it appeared failed, the
919        capturing was not being forgotten if a higher numbered group was later
920        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
921        group 1 to "a", when in fact it should be unset. This applied to multi-
922        branched capturing and non-capturing groups, repeated or not, and also to
923        positive assertions (capturing in negative assertions does not happen
924        in PCRE) and also to nested atomic groups.
925    
926    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
927        subject after a captured substring, to make it easier to tell which of a
928        number of identical substrings has been captured.
929    
930    24. The way atomic groups are processed by pcre_exec() has been changed so that
931        if they are repeated, backtracking one repetition now resets captured
932        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
933        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
934        "aaa". Previously, it would have been "a". As part of this code
935        refactoring, the way recursive calls are handled has also been changed.
936    
937    25. If an assertion condition captured any substrings, they were not passed
938        back unless some other capturing happened later. For example, if
939        (?(?=(a))a) was matched against "a", no capturing was returned.
940    
941    26. When studying a pattern that contained subroutine calls or assertions,
942        the code for finding the minimum length of a possible match was handling
943        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
944        group 1 called group 2 while simultaneously a separate group 2 called group
945        1). A stack overflow occurred in this case. I have fixed this by limiting
946        the recursion depth to 10.
947    
948    27. Updated RunTest.bat in the distribution to the version supplied by Tom
949        Fortmann. This supports explicit test numbers on the command line, and has
950        argument validation and error reporting.
951    
952    28. An instance of \X with an unlimited repeat could fail if at any point the
953        first character it looked at was a mark character.
954    
955    29. Some minor code refactoring concerning Unicode properties and scripts
956        should reduce the stack requirement of match() slightly.
957    
958    30. Added the '=' option to pcretest to check the setting of unused capturing
959        slots at the end of the pattern, which are documented as being -1, but are
960        not included in the return count.
961    
962    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
963        compiled something random. Now it gives a compile-time error (as does
964        Perl).
965    
966    32. A *MARK encountered during the processing of a positive assertion is now
967        recorded and passed back (compatible with Perl).
968    
969    33. If --only-matching or --colour was set on a pcregrep call whose pattern
970        had alternative anchored branches, the search for a second match in a line
971        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
972        matched the line "0102" twice. The same bug affected patterns that started
973        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
974        twice.
975    
976    34. Previously, PCRE did not allow quantification of assertions. However, Perl
977        does, and because of capturing effects, quantifying parenthesized
978        assertions may at times be useful. Quantifiers are now allowed for
979        parenthesized assertions.
980    
981    35. A minor code tidy in pcre_compile() when checking options for \R usage.
982    
983    36. \g was being checked for fancy things in a character class, when it should
984        just be a literal "g".
985    
986    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
987        appearance of a nested POSIX class supersedes an apparent external class.
988        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
989        unescaped square brackets may also appear as part of class names. For
990        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
991        more like Perl. (But see 8.20/1 above.)
992    
993    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
994        was because it thought it was \N{name}, which is not supported).
995    
996    39. Add minix to OS list not supporting the -S option in pcretest.
997    
998    40. PCRE tries to detect cases of infinite recursion at compile time, but it
999        cannot analyze patterns in sufficient detail to catch mutual recursions
1000        such as ((?1))((?2)). There is now a runtime test that gives an error if a
1001        subgroup is called recursively as a subpattern for a second time at the
1002        same position in the subject string. In previous releases this might have
1003        been caught by the recursion limit, or it might have run out of stack.
1004    
1005    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1006        happen only once. PCRE was, however incorrectly giving a compile time error
1007        "recursive call could loop indefinitely" because it cannot analyze the
1008        pattern in sufficient detail. The compile time test no longer happens when
1009        PCRE is compiling a conditional subpattern, but actual runaway loops are
1010        now caught at runtime (see 40 above).
1011    
1012    42. It seems that Perl allows any characters other than a closing parenthesis
1013        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1014        has been changed to be the same.
1015    
1016    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1017        as not to get warnings when autogen.sh is called. Also changed
1018        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1019    
1020    44. To help people who use pcregrep to scan files containing exceedingly long
1021        lines, the following changes have been made:
1022    
1023        (a) The default value of the buffer size parameter has been increased from
1024            8K to 20K. (The actual buffer used is three times this size.)
1025    
1026        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1027            PCRE is built.
1028    
1029        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1030            to be set at run time.
1031    
1032        (d) Numerical values in pcregrep options can be followed by K or M, for
1033            example --buffer-size=50K.
1034    
1035        (e) If a line being scanned overflows pcregrep's buffer, an error is now
1036            given and the return code is set to 2.
1037    
1038    45. Add a pointer to the latest mark to the callout data block.
1039    
1040    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1041        partial match of an empty string instead of no match. This was specific to
1042        the use of ".".
1043    
1044    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1045        complete match instead of a partial match. This bug was dependent on both
1046        the PCRE_UTF8 and PCRE_DOTALL options being set.
1047    
1048    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1049        starting byte set, because \b was not being ignored.
1050    
1051    
1052    Version 8.12 15-Jan-2011
1053    ------------------------
1054    
1055    1.  Fixed some typos in the markup of the man pages, and wrote a script that
1056        checks for such things as part of the documentation building process.
1057    
1058    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
1059        --match-limit and --recursion-limit options (added for 8.11). In
1060        particular, this made one of the standard tests fail. (The integer value
1061        went into the wrong half of a long int.)
1062    
1063    3.  If the --colour option was given to pcregrep with -v (invert match), it
1064        did strange things, either producing crazy output, or crashing. It should,
1065        of course, ignore a request for colour when reporting lines that do not
1066        match.
1067    
1068    4.  Another pcregrep bug caused similar problems if --colour was specified with
1069        -M (multiline) and the pattern match finished with a line ending.
1070    
1071    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
1072        matched in multiline mode, the following line was shown as part of the
1073        match. This seems wrong, so I have changed it.
1074    
1075    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
1076        the check for further matches in the same line (so they could be coloured)
1077        to overrun the end of the current line. If another match was found, it was
1078        incorrectly shown (and then shown again when found in the next line).
1079    
1080    7.  If pcregrep was compiled under Windows, there was a reference to the
1081        function pcregrep_exit() before it was defined. I am assuming this was
1082        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1083        reported by a user. I've moved the definition above the reference.
1084    
1085    
1086    Version 8.11 10-Dec-2010
1087    ------------------------
1088    
1089    1.  (*THEN) was not working properly if there were untried alternatives prior
1090        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1091        backtracked to try for "b" instead of moving to the next alternative branch
1092        at the same level (in this case, to look for "c"). The Perl documentation
1093        is clear that when (*THEN) is backtracked onto, it goes to the "next
1094        alternative in the innermost enclosing group".
1095    
1096    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1097        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
1098        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1099        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1100        (*THEN).
1101    
1102    3.  If \s appeared in a character class, it removed the VT character from
1103        the class, even if it had been included by some previous item, for example
1104        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1105        of \s, but is part of the POSIX "space" class.)
1106    
1107    4.  A partial match never returns an empty string (because you can always
1108        match an empty string at the end of the subject); however the checking for
1109        an empty string was starting at the "start of match" point. This has been
1110        changed to the "earliest inspected character" point, because the returned
1111        data for a partial match starts at this character. This means that, for
1112        example, /(?<=abc)def/ gives a partial match for the subject "abc"
1113        (previously it gave "no match").
1114    
1115    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1116        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1117        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1118        has an implication that the given string is incomplete (because a partial
1119        match is preferred over a full match). For this reason, these items now
1120        give a partial match in this situation. [Aside: previously, the one case
1121        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1122        match rather than a full match, which was wrong by the old rules, but is
1123        now correct.]
1124    
1125    6.  There was a bug in the handling of #-introduced comments, recognized when
1126        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1127        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1128        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1129        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1130        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1131        places in pcre_compile().
1132    
1133    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
1134        comments when looking ahead for named forward references to subpatterns,
1135        the only newline sequence it recognized was NL. It now handles newlines
1136        according to the set newline convention.
1137    
1138    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1139        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1140        cater for a lack of strerror(). These oversights have been fixed.
1141    
1142    9.  Added --match-limit and --recursion-limit to pcregrep.
1143    
1144    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1145    
1146    11. When the -o option was used, pcregrep was setting a return code of 1, even
1147        when matches were found, and --line-buffered was not being honoured.
1148    
1149    12. Added an optional parentheses number to the -o and --only-matching options
1150        of pcregrep.
1151    
1152    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1153        can match an empty string. The code to do it in pcretest and pcredemo
1154        needed fixing:
1155    
1156        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1157            only one byte after an empty string match just before CRLF (this case
1158            just got forgotten; "any" and "anycrlf" were OK).
1159    
1160        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1161            mode when an empty string match preceded an ASCII character followed by
1162            a non-ASCII character. (The code for advancing by one character rather
1163            than one byte was nonsense.)
1164    
1165        (c) The pcredemo.c sample program did not have any code at all to handle
1166            the cases when CRLF is a valid newline sequence.
1167    
1168    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1169        as a starting offset was within the subject string. There is now a new
1170        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1171        negative or greater than the length of the string. In order to test this,
1172        pcretest is extended to allow the setting of negative starting offsets.
1173    
1174    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1175        starting offset points to the beginning of a UTF-8 character was
1176        unnecessarily clumsy. I tidied it up.
1177    
1178    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1179        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1180    
1181    17. Nobody had reported that the --include_dir option, which was added in
1182        release 7.7 should have been called --include-dir (hyphen, not underscore)
1183        for compatibility with GNU grep. I have changed it to --include-dir, but
1184        left --include_dir as an undocumented synonym, and the same for
1185        --exclude-dir, though that is not available in GNU grep, at least as of
1186        release 2.5.4.
1187    
1188    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1189        characters from a string of bytes) have been redefined so as not to use
1190        loops, in order to improve performance in some environments. At the same
1191        time, I abstracted some of the common code into auxiliary macros to save
1192        repetition (this should not affect the compiled code).
1193    
1194    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1195        compile-time error is now given if \c is not followed by an ASCII
1196        character, that is, a byte less than 128. (In EBCDIC mode, the code is
1197        different, and any byte value is allowed.)
1198    
1199    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1200        START_OPTIMIZE option, which is now allowed at compile time - but just
1201        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1202        to pcregrep and other applications that have no direct access to PCRE
1203        options. The new /Y option in pcretest sets this option when calling
1204        pcre_compile().
1205    
1206    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1207        back references. Groups containing recursive back references were forced to
1208        be atomic by that change, but in the case of named groups, the amount of
1209        memory required was incorrectly computed, leading to "Failed: internal
1210        error: code overflow". This has been fixed.
1211    
1212    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1213        pcretest.c, to avoid build problems in some Borland environments.
1214    
1215    
1216    Version 8.10 25-Jun-2010
1217    ------------------------
1218    
1219    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1220        THEN.
1221    
1222    2.  (*ACCEPT) was not working when inside an atomic group.
1223    
1224    3.  Inside a character class, \B is treated as a literal by default, but
1225        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1226        causes the error). The code is unchanged, but I tidied the documentation.
1227    
1228    4.  Inside a character class, PCRE always treated \R and \X as literals,
1229        whereas Perl faults them if its -w option is set. I have changed PCRE so
1230        that it faults them when PCRE_EXTRA is set.
1231    
1232    5.  Added support for \N, which always matches any character other than
1233        newline. (It is the same as "." when PCRE_DOTALL is not set.)
1234    
1235    6.  When compiling pcregrep with newer versions of gcc which may have
1236        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1237        declared with attribute warn_unused_result" were given. Just casting the
1238        result to (void) does not stop the warnings; a more elaborate fudge is
1239        needed. I've used a macro to implement this.
1240    
1241    7.  Minor change to pcretest.c to avoid a compiler warning.
1242    
1243    8.  Added four artifical Unicode properties to help with an option to make
1244        \s etc use properties (see next item). The new properties are: Xan
1245        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1246    
1247    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1248        use Unicode properties. (*UCP) at the start of a pattern can be used to set
1249        this option. Modified pcretest to add /W to test this facility. Added
1250        REG_UCP to make it available via the POSIX interface.
1251    
1252    10. Added --line-buffered to pcregrep.
1253    
1254    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1255        studied, and the match started with a letter with a code point greater than
1256        127 whose first byte was different to the first byte of the other case of
1257        the letter, the other case of this starting letter was not recognized
1258        (#976).
1259    
1260    12. If a pattern that was studied started with a repeated Unicode property
1261        test, for example, \p{Nd}+, there was the theoretical possibility of
1262        setting up an incorrect bitmap of starting bytes, but fortunately it could
1263        not have actually happened in practice until change 8 above was made (it
1264        added property types that matched character-matching opcodes).
1265    
1266    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1267        possible starting bytes for non-anchored patterns.
1268    
1269    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1270        \R, and also a number of cases that involve Unicode properties, both
1271        explicit and implicit when PCRE_UCP is set.
1272    
1273    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1274        input, it could crash or give wrong results if characters with values
1275        greater than 0xc0 were present in the subject string. (Detail: it assumed
1276        UTF-8 input when processing these items.)
1277    
1278    16. Added a lot of (int) casts to avoid compiler warnings in systems where
1279        size_t is 64-bit (#991).
1280    
1281    17. Added a check for running out of memory when PCRE is compiled with
1282        --disable-stack-for-recursion (#990).
1283    
1284    18. If the last data line in a file for pcretest does not have a newline on
1285        the end, a newline was missing in the output.
1286    
1287    19. The default pcre_chartables.c file recognizes only ASCII characters (values
1288        less than 128) in its various bitmaps. However, there is a facility for
1289        generating tables according to the current locale when PCRE is compiled. It
1290        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1291        characters, are recognized by isspace() and therefore were getting set in
1292        these tables, and indeed these tables seem to approximate to ISO 8859. This
1293        caused a problem in UTF-8 mode when pcre_study() was used to create a list
1294        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1295        which of course cannot start UTF-8 characters. I have changed the code so
1296        that only real ASCII characters (less than 128) and the correct starting
1297        bytes for UTF-8 encodings are set for characters greater than 127 when in
1298        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1299        altogether.)
1300    
1301    20. Added the /T option to pcretest so as to be able to run tests with non-
1302        standard character tables, thus making it possible to include the tests
1303        used for 19 above in the standard set of tests.
1304    
1305    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1306        reference to a subpattern the other side of a comment that contains an
1307        opening parenthesis caused either an internal compiling error, or a
1308        reference to the wrong subpattern.
1309    
1310    
1311    Version 8.02 19-Mar-2010
1312    ------------------------
1313    
1314    1.  The Unicode data tables have been updated to Unicode 5.2.0.
1315    
1316    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
1317        configured.
1318    
1319    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
1320        original author of that file, following a query about its status.
1321    
1322    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
1323        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1324    
1325    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1326        quantifier applied to a forward-referencing subroutine call, could compile
1327        incorrect code or give the error "internal error: previously-checked
1328        referenced subpattern not found".
1329    
1330    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1331        variables to point to external functions. For these systems, therefore,
1332        pcre_malloc etc. are now initialized to local functions that call the
1333        relevant global functions.
1334    
1335    7.  There were two entries missing in the vectors called coptable and poptable
1336        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1337        I've fixed the data, and added a kludgy way of testing at compile time that
1338        the lengths are correct (equal to the number of opcodes).
1339    
1340    8.  Following on from 7, I added a similar kludge to check the length of the
1341        eint vector in pcreposix.c.
1342    
1343    9.  Error texts for pcre_compile() are held as one long string to avoid too
1344        much relocation at load time. To find a text, the string is searched,
1345        counting zeros. There was no check for running off the end of the string,
1346        which could happen if a new error number was added without updating the
1347        string.
1348    
1349    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1350    
1351    11. \K was not working if it appeared in an atomic group or in a group that
1352        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1353        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1354        the assertion is positive, but not if it is negative.
1355    
1356    12. Change 11 fortuitously reduced the size of the stack frame used in the
1357        "match()" function of pcre_exec.c by one pointer. Forthcoming
1358        implementation of support for (*MARK) will need an extra pointer on the
1359        stack; I have reserved it now, so that the stack frame size does not
1360        decrease.
1361    
1362    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1363        item in branch that calls a recursion is a subroutine call - as in the
1364        second branch in the above example - was incorrectly given the compile-
1365        time error "recursive call could loop indefinitely" because pcre_compile()
1366        was not correctly checking the subroutine for matching a non-empty string.
1367    
1368    14. The checks for overrunning compiling workspace could trigger after an
1369        overrun had occurred. This is a "should never occur" error, but it can be
1370        triggered by pathological patterns such as hundreds of nested parentheses.
1371        The checks now trigger 100 bytes before the end of the workspace.
1372    
1373    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1374    
1375    
1376    Version 8.01 19-Jan-2010
1377    ------------------------
1378    
1379    1.  If a pattern contained a conditional subpattern with only one branch (in
1380        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1381        computed the wrong minimum data length (which is of course zero for such
1382        subpatterns). This could cause incorrect "no match" results.
1383    
1384    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1385        the pattern is reset in the first branch, pcre_compile() failed with
1386        "internal error: code overflow at offset...". This happened only when
1387        the reset was to the original external option setting. (An optimization
1388        abstracts leading options settings into an external setting, which was the
1389        cause of this.)
1390    
1391    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1392        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1393        assertion pattern did not match (meaning that the assertion was true), it
1394        was incorrectly treated as false if the SKIP had been reached during the
1395        matching. This also applied to assertions used as conditions.
1396    
1397    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1398        assertion subpattern, including such a pattern used as a condition,
1399        unpredictable results occurred, instead of the error return
1400        PCRE_ERROR_DFA_UITEM.
1401    
1402    5.  The C++ GlobalReplace function was not working like Perl for the special
1403        situation when an empty string is matched. It now does the fancy magic
1404        stuff that is necessary.
1405    
1406    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1407        removed. (These were left over from very, very early versions of PCRE.)
1408    
1409    7.  Some cosmetic changes to the code to make life easier when compiling it
1410        as part of something else:
1411    
1412        (a) Change DEBUG to PCRE_DEBUG.
1413    
1414        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1415            called "current" as "current_branch", to prevent a collision with the
1416            Linux macro when compiled as a kernel module.
1417    
1418        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1419            prevent a collision with the Linux macro when compiled as a kernel
1420            module.
1421    
1422    8.  In pcre_compile() there are some checks for integer overflows that used to
1423        cast potentially large values to (double). This has been changed to that
1424        when building, a check for int64_t is made, and if it is found, it is used
1425        instead, thus avoiding the use of floating point arithmetic. (There is no
1426        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1427        double.
1428    
1429    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1430        2005 (difference between two addresses compared to an unsigned value).
1431    
1432    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1433        custom one, because of the following reported problem in Windows:
1434    
1435          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1436              under Win32.
1437          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1438              therefore missing the function definition.
1439          - The compiler thus generates a "C" signature for the test function.
1440          - The linker fails to find the "C" function.
1441          - PCRE fails to configure if asked to do so against libbz2.
1442    
1443    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1444        messages were output:
1445    
1446          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1447          rerunning libtoolize, to keep the correct libtool macros in-tree.
1448          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1449    
1450        I have done both of these things.
1451    
1452    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1453        most of the time, it *can* run out if it is given a pattern that contains a
1454        runaway infinite recursion. I updated the discussion in the pcrestack man
1455        page.
1456    
1457    13. Now that we have gone to the x.xx style of version numbers, the minor
1458        version may start with zero. Using 08 or 09 is a bad idea because users
1459        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1460        interpreted as invalid octal numbers. I've updated the previous comment in
1461        configure.ac, and also added a check that gives an error if 08 or 09 are
1462        used.
1463    
1464    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1465        causing partial matching to fail when the end of the subject matched \W
1466        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1467    
1468    15. There were some discrepancies between the declarations in pcre_internal.h
1469        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1470        their definitions. The declarations used "const uschar *" and the
1471        definitions used USPTR. Even though USPTR is normally defined as "const
1472        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1473        reported that: "This difference in casting confuses some C++ compilers, for
1474        example, SunCC recognizes above declarations as different functions and
1475        generates broken code for hbpcre." I have changed the declarations to use
1476        USPTR.
1477    
1478    16. GNU libtool is named differently on some systems. The autogen.sh script now
1479        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1480        (FreeBSD).
1481    
1482    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1483        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1484        comment: "Figure out how to create a longlong from a string: strtoll and
1485        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1486        instance, but it only takes 2 args instead of 3!"
1487    
1488    18. A subtle bug concerned with back references has been fixed by a change of
1489        specification, with a corresponding code fix. A pattern such as
1490        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1491        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1492        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1493        same bug. Such groups have to be quantified to be useful, or contained
1494        inside another quantified group. (If there's no repetition, the reference
1495        can never match.) The problem arises because, having left the group and
1496        moved on to the rest of the pattern, a later failure that backtracks into
1497        the group uses the captured value from the final iteration of the group
1498        rather than the correct earlier one. I have fixed this in PCRE by forcing
1499        any group that contains a reference to itself to be an atomic group; that
1500        is, there cannot be any backtracking into it once it has completed. This is
1501        similar to recursive and subroutine calls.
1502    
1503    
1504    Version 8.00 19-Oct-09
1505    ----------------------
1506    
1507    1.  The table for translating pcre_compile() error codes into POSIX error codes
1508        was out-of-date, and there was no check on the pcre_compile() error code
1509        being within the table. This could lead to an OK return being given in
1510        error.
1511    
1512    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1513        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1514        in a Windows environment.
1515    
1516    3.  The pcregrep --count option prints the count for each file even when it is
1517        zero, as does GNU grep. However, pcregrep was also printing all files when
1518        --files-with-matches was added. Now, when both options are given, it prints
1519        counts only for those files that have at least one match. (GNU grep just
1520        prints the file name in this circumstance, but including the count seems
1521        more useful - otherwise, why use --count?) Also ensured that the
1522        combination -clh just lists non-zero counts, with no names.
1523    
1524    4.  The long form of the pcregrep -F option was incorrectly implemented as
1525        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1526        but it seems right to fix it, and I didn't think it was worth preserving
1527        the old behaviour.
1528    
1529    5.  The command line items --regex=pattern and --regexp=pattern were not
1530        recognized by pcregrep, which required --regex pattern or --regexp pattern
1531        (with a space rather than an '='). The man page documented the '=' forms,
1532        which are compatible with GNU grep; these now work.
1533    
1534    6.  No libpcreposix.pc file was created for pkg-config; there was just
1535        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1536    
1537    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1538        when UCP support is not needed, by modifying the Python script that
1539        generates it from Unicode data files. This should not matter if the module
1540        is correctly used as a library, but I received one complaint about 50K of
1541        unwanted data. My guess is that the person linked everything into his
1542        program rather than using a library. Anyway, it does no harm.
1543    
1544    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1545        was a minimum greater than 1 for a wide character in a possessive
1546        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1547        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1548        character. Chaos in the form of incorrect output or a compiling loop could
1549        result.
1550    
1551    9.  The restrictions on what a pattern can contain when partial matching is
1552        requested for pcre_exec() have been removed. All patterns can now be
1553        partially matched by this function. In addition, if there are at least two
1554        slots in the offset vector, the offset of the earliest inspected character
1555        for the match and the offset of the end of the subject are set in them when
1556        PCRE_ERROR_PARTIAL is returned.
1557    
1558    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1559        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1560        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1561        and may be more useful for multi-segment matching.
1562    
1563    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1564        used to be given if ever the end of the subject was reached; now it is
1565        given only if matching could not proceed because another character was
1566        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1567        string "Z", which now yields "no match" instead of "partial match". In the
1568        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1569        final character ended with (*FAIL).
1570    
1571    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1572        if the pattern had a "must contain" character that was already found in the
1573        earlier partial match, unless partial matching was again requested. For
1574        example, with the pattern /dog.(body)?/, the "must contain" character is
1575        "g". If the first part-match was for the string "dog", restarting with
1576        "sbody" failed. This bug has been fixed.
1577    
1578    13. The string returned by pcre_dfa_exec() after a partial match has been
1579        changed so that it starts at the first inspected character rather than the
1580        first character of the match. This makes a difference only if the pattern
1581        starts with a lookbehind assertion or \b or \B (\K is not supported by
1582        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1583        matching functions compatible, and I think it's the right thing to do.
1584    
1585    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1586        so that the demonstration program is easily available in environments where
1587        PCRE has not been installed from source.
1588    
1589    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1590        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1591        library.
1592    
1593    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1594        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1595        is not the first non-POSIX option to be added. Clearly some people find
1596        these options useful.
1597    
1598    17. If a caller to the POSIX matching function regexec() passes a non-zero
1599        value for nmatch with a NULL value for pmatch, the value of
1600        nmatch is forced to zero.
1601    
1602    18. RunGrepTest did not have a test for the availability of the -u option of
1603        the diff command, as RunTest does. It now checks in the same way as
1604        RunTest, and also checks for the -b option.
1605    
1606    19. If an odd number of negated classes containing just a single character
1607        interposed, within parentheses, between a forward reference to a named
1608        subpattern and the definition of the subpattern, compilation crashed with
1609        an internal error, complaining that it could not find the referenced
1610        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1611        [The bug was that it was starting one character too far in when skipping
1612        over the character class, thus treating the ] as data rather than
1613        terminating the class. This meant it could skip too much.]
1614    
1615    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1616        /g option in pcretest when the pattern contains \K, which makes it possible
1617        to have an empty string match not at the start, even when the pattern is
1618        anchored. Updated pcretest and pcredemo to use this option.
1619    
1620    21. If the maximum number of capturing subpatterns in a recursion was greater
1621        than the maximum at the outer level, the higher number was returned, but
1622        with unset values at the outer level. The correct (outer level) value is
1623        now given.
1624    
1625    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1626        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1627        make it do so. The string so far is captured, making this feature
1628        compatible with Perl.
1629    
1630    23. The tests have been re-organized, adding tests 11 and 12, to make it
1631        possible to check the Perl 5.10 features against Perl 5.10.
1632    
1633    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1634        pattern matches a fixed length string. PCRE did not allow this; now it
1635        does. Neither allows recursion.
1636    
1637    25. I finally figured out how to implement a request to provide the minimum
1638        length of subject string that was needed in order to match a given pattern.
1639        (It was back references and recursion that I had previously got hung up
1640        on.) This code has now been added to pcre_study(); it finds a lower bound
1641        to the length of subject needed. It is not necessarily the greatest lower
1642        bound, but using it to avoid searching strings that are too short does give
1643        some useful speed-ups. The value is available to calling programs via
1644        pcre_fullinfo().
1645    
1646    26. While implementing 25, I discovered to my embarrassment that pcretest had
1647        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1648        study optimizations had never been tested with that matching function.
1649        Oops. What is worse, even when it was passed study data, there was a bug in
1650        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1651        were also very few tests of studied patterns with pcre_dfa_exec().
1652    
1653    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1654        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1655        on the other side of the coin, they are no longer allowed to have different
1656        names, because these cannot be distinguished in PCRE, and this has caused
1657        confusion. (This is a difference from Perl.)
1658    
1659    28. When duplicate subpattern names are present (necessarily with different
1660        numbers, as required by 27 above), and a test is made by name in a
1661        conditional pattern, either for a subpattern having been matched, or for
1662        recursion in such a pattern, all the associated numbered subpatterns are
1663        tested, and the overall condition is true if the condition is true for any
1664        one of them. This is the way Perl works, and is also more like the way
1665        testing by number works.
1666    
1667    
1668    Version 7.9 11-Apr-09
1669    ---------------------
1670    
1671    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1672        (pcretest), all targets were linked against these libraries. This included
1673        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1674        libraries. This caused unwanted dependencies to be created. This problem
1675        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1676        pcretest is linked with readline.
1677    
1678    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1679        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1680        moved outside it again, because FALSE and TRUE are already defined in AIX,
1681        but BOOL is not.
1682    
1683    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1684        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1685    
1686    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1687        hyphens) following file names and line numbers when outputting matching
1688        lines. This is not true; no spaces are inserted. I have also clarified the
1689        wording for the --colour (or --color) option.
1690    
1691    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1692        was not coloured; this is different to GNU grep, so I have changed it to be
1693        the same.
1694    
1695    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1696        each matching line was coloured. Now it goes on to look for further matches
1697        of any of the test patterns, which is the same behaviour as GNU grep.
1698    
1699    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1700        doesn't make sense to accept an empty string match in pcregrep, so I have
1701        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1702        seems to be how GNU grep behaves. [But see later change 40 for release
1703        8.33.]
1704    
1705    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1706        start or after a newline", because the conditional assertion was not being
1707        correctly handled. The rule now is that both the assertion and what follows
1708        in the first alternative must satisfy the test.
1709    
1710    9.  If auto-callout was enabled in a pattern with a conditional group whose
1711        condition was an assertion, PCRE could crash during matching, both with
1712        pcre_exec() and pcre_dfa_exec().
1713    
1714    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1715        used for matching.
1716    
1717    11. Unicode property support in character classes was not working for
1718        characters (bytes) greater than 127 when not in UTF-8 mode.
1719    
1720    12. Added the -M command line option to pcretest.
1721    
1722    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1723    
1724    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1725    
1726    16. Added comments and documentation about mis-use of no_arg in the C++
1727        wrapper.
1728    
1729    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1730        from Martin Jerabek that uses macro names for all relevant character and
1731        string constants.
1732    
1733    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1734        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1735        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1736        these, but not everybody uses configure.
1737    
1738    19. A conditional group that had only one branch was not being correctly
1739        recognized as an item that could match an empty string. This meant that an
1740        enclosing group might also not be so recognized, causing infinite looping
1741        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1742        with the subject "ab", where knowledge that the repeated group can match
1743        nothing is needed in order to break the loop.
1744    
1745    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1746        exec(), but without supplying a callout function, matching went wrong.
1747    
1748    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1749        leak if the size of the offset vector was greater than 30. When the vector
1750        is smaller, the saved offsets during recursion go onto a local stack
1751        vector, but for larger vectors malloc() is used. It was failing to free
1752        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1753        error, in fact).
1754    
1755    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1756        heapframe that is used only when UTF-8 support is enabled. This caused no
1757        problem, but was untidy.
1758    
1759    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1760        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1761        included within another project.
1762    
1763    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1764        slightly modified by me:
1765    
1766          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1767              not building pcregrep.
1768    
1769          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1770              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1771    
1772    25. Forward references, both numeric and by name, in patterns that made use of
1773        duplicate group numbers, could behave incorrectly or give incorrect errors,
1774        because when scanning forward to find the reference group, PCRE was not
1775        taking into account the duplicate group numbers. A pattern such as
1776        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1777    
1778    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1779        the feature of a custom pointer more persuasive (as requested by a user).
1780    
1781    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1782        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1783        pre-defined.
1784    
1785    28. Added support for (*UTF8) at the start of a pattern.
1786    
1787    29. Arrange for flags added by the "release type" setting in CMake to be shown
1788        in the configuration summary.
1789    
1790    
1791    Version 7.8 05-Sep-08
1792    ---------------------
1793    
1794    1.  Replaced UCP searching code with optimized version as implemented for Ad
1795        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1796        stage table and inline lookup instead of a function, giving speed ups of 2
1797        to 5 times on some simple patterns that I tested. Permission was given to
1798        distribute the MultiStage2.py script that generates the tables (it's not in
1799        the tarball, but is in the Subversion repository).
1800    
1801    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1802        scripts.
1803    
1804    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1805        a group with a zero qualifier. The result of the study could be incorrect,
1806        or the function might crash, depending on the pattern.
1807    
1808    4.  Caseless matching was not working for non-ASCII characters in back
1809        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1810        It now works when Unicode Property Support is available.
1811    
1812    5.  In pcretest, an escape such as \x{de} in the data was always generating
1813        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1814        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1815        truncation.
1816    
1817    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1818    
1819    7.  Added two (int) casts to pcregrep when printing the difference of two
1820        pointers, in case they are 64-bit values.
1821    
1822    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1823        test 2 if it fails.
1824    
1825    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1826        and a #define of that name to empty if it is not externally set. This is to
1827        allow users of MSVC to set it if necessary.
1828    
1829    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1830        the convenience functions in the pcre_get.c source file.
1831    
1832    11. An option change at the start of a pattern that had top-level alternatives
1833        could cause overwriting and/or a crash. This command provoked a crash in
1834        some environments:
1835    
1836          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1837    
1838        This potential security problem was recorded as CVE-2008-2371.
1839    
1840    12. For a pattern where the match had to start at the beginning or immediately
1841        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1842        pcre_dfa_exec() could read past the end of the passed subject if there was
1843        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1844        pcretest so that it places the subject at the end of its malloc-ed buffer.
1845    
1846    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1847        exec() might read past the end of the data buffer in UTF-8 mode.
1848    
1849    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1850        the data contained the byte 0x85 as part of a UTF-8 character within its
1851        first line. This applied both to normal and DFA matching.
1852    
1853    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1854        /^[^d]*?$/8 failed to match "abc".
1855    
1856    16. Added a missing copyright notice to pcrecpp_internal.h.
1857    
1858    17. Make it more clear in the documentation that values returned from
1859        pcre_exec() in ovector are byte offsets, not character counts.
1860    
1861    18. Tidied a few places to stop certain compilers from issuing warnings.
1862    
1863    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1864        supplied by Stefan Weber. I made a further small update for 7.8 because
1865        there is a change of source arrangements: the pcre_searchfuncs.c module is
1866        replaced by pcre_ucd.c.
1867    
1868    
1869    Version 7.7 07-May-08
1870    ---------------------
1871    
1872    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1873        a string to a long long, pretend we don't even have a long long." This is
1874        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1875    
1876    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1877        pre-7.6 versions, which defined a global no_arg variable instead of putting
1878        it in the RE class. (See also #8 below.)
1879    
1880    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1881        Lopes.
1882    
1883    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1884    
1885        (1) The include/exclude patterns were being applied to the whole pathnames
1886            of files, instead of just to the final components.
1887    
1888        (2) If there was more than one level of directory, the subdirectories were
1889            skipped unless they satisfied the include/exclude conditions. This is
1890            inconsistent with GNU grep (and could even be seen as contrary to the
1891            pcregrep specification - which I improved to make it absolutely clear).
1892            The action now is always to scan all levels of directory, and just
1893            apply the include/exclude patterns to regular files.
1894    
1895    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1896        --exclude_dir in the tests to avoid scanning .svn directories.
1897    
1898    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1899        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1900        doesn't support NULs in patterns.
1901    
1902    7.  Added some missing "const"s to declarations of static tables in
1903        pcre_compile.c and pcre_dfa_exec.c.
1904    
1905    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1906        caused by fix #2  above. (Subsequently also a second patch to fix the
1907        first patch. And a third patch - this was a messy problem.)
1908    
1909    9.  Applied Craig's patch to remove the use of push_back().
1910    
1911    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1912        matching function regexec().
1913    
1914    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1915        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1916        references. PCRE supports relative numbers with this syntax (I don't think
1917        Oniguruma does).
1918    
1919    12. Previously, a group with a zero repeat such as (...){0} was completely
1920        omitted from the compiled regex. However, this means that if the group
1921        was called as a subroutine from elsewhere in the pattern, things went wrong
1922        (an internal error was given). Such groups are now left in the compiled
1923        pattern, with a new opcode that causes them to be skipped at execution
1924        time.
1925    
1926    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1927        to the way PCRE behaves:
1928    
1929        (a) A lone ] character is dis-allowed (Perl treats it as data).
1930    
1931        (b) A back reference to an unmatched subpattern matches an empty string
1932            (Perl fails the current match path).
1933    
1934        (c) A data ] in a character class must be notated as \] because if the
1935            first data character in a class is ], it defines an empty class. (In
1936            Perl it is not possible to have an empty class.) The empty class []
1937            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1938            The negative empty class [^] matches any one character, independently
1939            of the DOTALL setting.
1940    
1941    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1942        non-existent subpattern following a character class starting with ']' and
1943        containing () gave an internal compiling error instead of "reference to
1944        non-existent subpattern". Fortunately, when the pattern did exist, the
1945        compiled code was correct. (When scanning forwards to check for the
1946        existence of the subpattern, it was treating the data ']' as terminating
1947        the class, so got the count wrong. When actually compiling, the reference
1948        was subsequently set up correctly.)
1949    
1950    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1951        it was being rejected as not supported by pcre_dfa_exec(), even though
1952        other assertions are supported. I have made pcre_dfa_exec() support
1953        (*FAIL).
1954    
1955    16. The implementation of 13c above involved the invention of a new opcode,
1956        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1957        cannot be changed at match time, I realized I could make a small
1958        improvement to matching performance by compiling OP_ALLANY instead of
1959        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1960        on the OP_ANY path.
1961    
1962    17. Compiling pcretest on Windows with readline support failed without the
1963        following two fixes: (1) Make the unistd.h include conditional on
1964        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1965    
1966    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1967        ncurses library to be included for pcretest when ReadLine support is
1968        requested, but also to allow for it to be overridden. This patch came from
1969        Daniel Bergström.
1970    
1971    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1972        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1973        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1974        spotting this.
1975    
1976    
1977    Version 7.6 28-Jan-08
1978    ---------------------
1979    
1980    1.  A character class containing a very large number of characters with
1981        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1982        overflow.
1983    
1984    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1985        HAVE_LONG_LONG is not defined.
1986    
1987    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1988        bring them up to date and include new features. This patch includes:
1989    
1990        - Fixed PH's badly added libz and libbz2 support.
1991        - Fixed a problem with static linking.
1992        - Added pcredemo. [But later removed - see 7 below.]
1993        - Fixed dftables problem and added an option.
1994        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1995            HAVE_LONG_LONG.
1996        - Added readline support for pcretest.
1997        - Added an listing of the option settings after cmake has run.
1998    
1999    4.  A user submitted a patch to Makefile that makes it easy to create
2000        "pcre.dll" under mingw when using Configure/Make. I added stuff to
2001        Makefile.am that cause it to include this special target, without
2002        affecting anything else. Note that the same mingw target plus all
2003        the other distribution libraries and programs are now supported
2004        when configuring with CMake (see 6 below) instead of with
2005        Configure/Make.
2006    
2007    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2008        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2009        exported in the Windows port". It has not yet been confirmed that the patch
2010        solves the problem, but it does no harm.
2011    
2012    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2013        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2014        with CMake, and also correct the comment about stack recursion.
2015    
2016    7.  Remove the automatic building of pcredemo from the ./configure system and
2017        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2018        of a program that users should build themselves after PCRE is installed, so
2019        building it automatically is not really right. What is more, it gave
2020        trouble in some build environments.
2021    
2022    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
2023    
2024    
2025    Version 7.5 10-Jan-08
2026    ---------------------
2027    
2028    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
2029        values in parens when parsing an RE using the C++ wrapper."
2030    
2031    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
2032        Characters greater than 255 were excluded from the class instead of being
2033        included.
2034    
2035    3.  The same bug as (2) above applied to negated POSIX classes such as
2036        [:^space:].
2037    
2038    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2039        defined or documented. It seems to have been a typo for PCRE_STATIC, so
2040        I have changed it.
2041    
2042    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
2043        first named subpattern) and a construct such as (?&a) would reference the
2044        first named subpattern whose name started with "a" (in other words, the
2045        length check was missing). Both these problems are fixed. "Subpattern name
2046        expected" is now given for (?&) (a zero-length name), and this patch also
2047        makes it give the same error for \k'' (previously it complained that that
2048        was a reference to a non-existent subpattern).
2049    
2050    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
2051        this is right because (?- can be followed by option settings as well as by
2052        digits. I have, however, made the messages clearer.
2053    
2054    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2055        than the number used in the conditional) now cause a compile-time error.
2056        This is actually not compatible with Perl, which accepts such patterns, but
2057        treats the conditional as always being FALSE (as PCRE used to), but it
2058        seems to me that giving a diagnostic is better.
2059    
2060    8.  Change "alphameric" to the more common word "alphanumeric" in comments
2061        and messages.
2062    
2063    9.  Fix two occurrences of "backslash" in comments that should have been
2064        "backspace".
2065    
2066    10. Remove two redundant lines of code that can never be obeyed (their function
2067        was moved elsewhere).
2068    
2069    11. The program that makes PCRE's Unicode character property table had a bug
2070        which caused it to generate incorrect table entries for sequences of
2071        characters that have the same character type, but are in different scripts.
2072        It amalgamated them into a single range, with the script of the first of
2073        them. In other words, some characters were in the wrong script. There were
2074        thirteen such cases, affecting characters in the following ranges:
2075    
2076          U+002b0 - U+002c1
2077          U+0060c - U+0060d
2078          U+0061e - U+00612
2079          U+0064b - U+0065e
2080          U+0074d - U+0076d
2081          U+01800 - U+01805
2082          U+01d00 - U+01d77
2083          U+01d9b - U+01dbf
2084          U+0200b - U+0200f
2085          U+030fc - U+030fe
2086          U+03260 - U+0327f
2087          U+0fb46 - U+0fbb1
2088          U+10450 - U+1049d
2089    
2090    12. The -o option (show only the matching part of a line) for pcregrep was not
2091        compatible with GNU grep in that, if there was more than one match in a
2092        line, it showed only the first of them. It now behaves in the same way as
2093        GNU grep.
2094    
2095    13. If the -o and -v options were combined for pcregrep, it printed a blank
2096        line for every non-matching line. GNU grep prints nothing, and pcregrep now
2097        does the same. The return code can be used to tell if there were any
2098        non-matching lines.
2099    
2100    14. Added --file-offsets and --line-offsets to pcregrep.
2101    
2102    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2103        infinitely looping recursion. The bug was that positive lookaheads were not
2104        being skipped when checking for a possible empty match (negative lookaheads
2105        and both kinds of lookbehind were skipped).
2106    
2107    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2108        inclusion of <windows.h> to before rather than after the definition of
2109        INVALID_FILE_ATTRIBUTES (patch from David Byron).
2110    
2111    17. Specifying a possessive quantifier with a specific limit for a Unicode
2112        character property caused pcre_compile() to compile bad code, which led at
2113        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2114        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2115        caused the error; without that there was no problem.
2116    
2117    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2118    
2119    19. Added --enable-pcretest-libreadline.
2120    
2121    20. In pcrecpp.cc, the variable 'count' was incremented twice in
2122        RE::GlobalReplace(). As a result, the number of replacements returned was
2123        double what it should be. I removed one of the increments, but Craig sent a
2124        later patch that removed the other one (the right fix) and added unit tests
2125        that check the return values (which was not done before).
2126    
2127    21. Several CMake things:
2128    
2129        (1) Arranged that, when cmake is used on Unix, the libraries end up with
2130            the names libpcre and libpcreposix, not just pcre and pcreposix.
2131    
2132        (2) The above change means that pcretest and pcregrep are now correctly
2133            linked with the newly-built libraries, not previously installed ones.
2134    
2135        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2136    
2137    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2138        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2139        UTF-8 newline character). The key issue is that the pattern starts .*;
2140        this means that the match must be either at the beginning, or after a
2141        newline. The bug was in the code for advancing after a failed match and
2142        checking that the new position followed a newline. It was not taking
2143        account of UTF-8 characters correctly.
2144    
2145    23. PCRE was behaving differently from Perl in the way it recognized POSIX
2146        character classes. PCRE was not treating the sequence [:...:] as a
2147        character class unless the ... were all letters. Perl, however, seems to
2148        allow any characters between [: and :], though of course it rejects as
2149        unknown any "names" that contain non-letters, because all the known class
2150        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2151        for example, whereas PCRE did not - it did not recognize a POSIX character
2152        class. This seemed a bit dangerous, so the code has been changed to be
2153        closer to Perl. The behaviour is not identical to Perl, because PCRE will
2154        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2155        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2156        Perl does, and where it didn't before.
2157    
2158    24. Rewrite so as to remove the single use of %n from pcregrep because in some
2159        Windows environments %n is disabled by default.
2160    
2161    
2162    Version 7.4 21-Sep-07
2163    ---------------------
2164    
2165    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
2166        means that a class such as [\s] counted as "explicit reference to CR or
2167        LF". That isn't really right - the whole point of the change was to try to
2168        help when there was an actual mention of one of the two characters. So now
2169        the change happens only if \r or \n (or a literal CR or LF) character is
2170        encountered.
2171    
2172    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
2173        of both had grown to the point where there were only 3 bits left.
2174        Fortunately, there was spare space in the data structure, and so I have
2175        moved the internal flags into a new 16-bit field to free up more option
2176        bits.
2177    
2178    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2179        but did not set the internal JCHANGED flag - either of these is enough to
2180        control the way the "get" function works - but the PCRE_INFO_JCHANGED
2181        facility is supposed to tell if (?J) was ever used, so now (?J) at the
2182        start sets both bits.
2183    
2184    4.  Added options (at build time, compile time, exec time) to change \R from
2185        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2186    
2187    5.  doc/pcresyntax.html was missing from the distribution.
2188    
2189    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2190        compatibility, even though it is no longer used.
2191    
2192    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2193        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2194        windows.h file is present (where different names are used). [This was
2195        reversed later after testing - see 16 below.]
2196    
2197    8.  Changed all #include <config.h> to #include "config.h". There were also
2198        some further <pcre.h> cases that I changed to "pcre.h".
2199    
2200    9.  When pcregrep was used with the --colour option, it missed the line ending
2201        sequence off the lines that it output.
2202    
2203    10. It was pointed out to me that arrays of string pointers cause lots of
2204        relocations when a shared library is dynamically loaded. A technique of
2205        using a single long string with a table of offsets can drastically reduce
2206        these. I have refactored PCRE in four places to do this. The result is
2207        dramatic:
2208    
2209          Originally:                          290
2210          After changing UCP table:            187
2211          After changing error message table:   43
2212          After changing table of "verbs"       36
2213          After changing table of Posix names   22
2214    
2215        Thanks to the folks working on Gregex for glib for this insight.
2216    
2217    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2218        unicode-properties was also set.
2219    
2220    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2221    
2222    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2223        checked only for CRLF.
2224    
2225    14. Added casts to pcretest.c to avoid compiler warnings.
2226    
2227    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2228    
2229    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2230        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2231        entirely. This removes changes made in 7 above.
2232    
2233    17. The CMake files have been updated, and there is now more information about
2234        building with CMake in the NON-UNIX-USE document.
2235    
2236    
2237    Version 7.3 28-Aug-07
2238    ---------------------
2239    
2240     1. In the rejigging of the build system that eventually resulted in 7.1, the
2241        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2242        brackets there is not right, since it causes compilers to look for an
2243        installed pcre.h, not the version that is in the source that is being
2244        compiled (which of course may be different). I have changed it back to:
2245    
2246          #include "pcre.h"
2247    
2248        I have a vague recollection that the change was concerned with compiling in
2249        different directories, but in the new build system, that is taken care of
2250        by the VPATH setting the Makefile.
2251    
2252     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2253        when the subject happened to end in the byte 0x85 (e.g. if the last
2254        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2255        characters but of course it shouldn't be taken as a newline when it is part
2256        of another character. The bug was that, for an unlimited repeat of . in
2257        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2258        characters when looking for a newline.
2259    
2260     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2261    
2262     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2263        in debug output.
2264    
2265     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2266        long printing in the pcrecpp unittest when running under MinGW.
2267    
2268     6. ESC_K was left out of the EBCDIC table.
2269    
2270     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2271        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2272        limit also applies to "virtual nesting" when a pattern is recursive, and in
2273        this case 1000 isn't so big. I have been able to remove this limit at the
2274        expense of backing off one optimization in certain circumstances. Normally,
2275        when pcre_exec() would call its internal match() function recursively and
2276        immediately return the result unconditionally, it uses a "tail recursion"
2277        feature to save stack. However, when a subpattern that can match an empty
2278        string has an unlimited repetition quantifier, it no longer makes this
2279        optimization. That gives it a stack frame in which to save the data for
2280        checking that an empty string has been matched. Previously this was taken
2281        from the 1000-entry workspace that had been reserved. So now there is no
2282        explicit limit, but more stack is used.
2283    
2284     8. Applied Daniel's patches to solve problems with the import/export magic
2285        syntax that is required for Windows, and which was going wrong for the
2286        pcreposix and pcrecpp parts of the library. These were overlooked when this
2287        problem was solved for the main library.
2288    
2289     9. There were some crude static tests to avoid integer overflow when computing
2290        the size of patterns that contain repeated groups with explicit upper
2291        limits. As the maximum quantifier is 65535, the maximum group length was
2292        set at 30,000 so that the product of these two numbers did not overflow a
2293        32-bit integer. However, it turns out that people want to use groups that
2294        are longer than 30,000 bytes (though not repeat them that many times).
2295        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2296        made it possible to implement the integer overflow checks in a much more
2297        dynamic way, which I have now done. The artificial limitation on group
2298        length has been removed - we now have only the limit on the total length of
2299        the compiled pattern, which depends on the LINK_SIZE setting.
2300    
2301    10. Fixed a bug in the documentation for get/copy named substring when
2302        duplicate names are permitted. If none of the named substrings are set, the
2303        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2304        empty string.
2305    
2306    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2307        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2308        because the ] is interpreted as the first data character and the
2309        terminating ] is not found. PCRE has been made compatible with Perl in this
2310        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2311        cause memory overwriting.
2312    
2313    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2314        string has been matched (to stop an infinite loop). It was not recognizing
2315        a conditional subpattern that could match an empty string if that
2316        subpattern was within another subpattern. For example, it looped when
2317        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
2318        condition was not nested. This bug has been fixed.
2319    
2320    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2321        past the start of the subject in the presence of bytes with the top bit
2322        set, for example "\x8aBCD".
2323    
2324    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2325        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2326    
2327    14. Optimized (?!) to (*FAIL).
2328    
2329    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2330        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2331        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2332        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2333        does: it's just the validity check that is more restrictive.
2334    
2335    16. Inserted checks for integer overflows during escape sequence (backslash)
2336        processing, and also fixed erroneous offset values for syntax errors during
2337        backslash processing.
2338    
2339    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2340        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2341    
2342    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2343        caused an overrun.
2344    
2345    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2346        something other than just ASCII characters) inside a group that had an
2347        unlimited repeat caused a loop at compile time (while checking to see
2348        whether the group could match an empty string).
2349    
2350    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2351        [\P{Any}] did so. (Error in the code for printing property names.)
2352    
2353    21. An orphan \E inside a character class could cause a crash.
2354    
2355    22. A repeated capturing bracket such as (A)? could cause a wild memory
2356        reference during compilation.
2357    
2358    23. There are several functions in pcre_compile() that scan along a compiled
2359        expression for various reasons (e.g. to see if it's fixed length for look
2360        behind). There were bugs in these functions when a repeated \p or \P was
2361        present in the pattern. These operators have additional parameters compared
2362        with \d, etc, and these were not being taken into account when moving along
2363        the compiled data. Specifically:
2364    
2365        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2366            length.
2367    
2368        (b) An item such as \pL+ within a repeated group could cause crashes or
2369            loops.
2370    
2371        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2372            "reference to non-existent subpattern" error.
2373    
2374        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2375    
2376    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2377        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2378    
2379    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2380    
2381    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2382        character were causing crashes (broken optimization).
2383    
2384    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2385        \p or \P) caused a compile-time loop.
2386    
2387    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2388        break. For example, the unstudied pattern [\r\n]A does not match the string
2389        "\r\nA" because change 7.0/46 below moves the current point on by two
2390        characters after failing to match at the start. However, the pattern \nA
2391        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2392        the same is true. There doesn't seem any very clean way out of this, but
2393        what I have chosen to do makes the common cases work: PCRE now takes note
2394        of whether there can be an explicit match for \r or \n anywhere in the
2395        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2396        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2397        pattern has explicit CR or LF references.
2398    
2399    29. Added (*CR) etc for changing newline setting at start of pattern.
2400    
2401    
2402    Version 7.2 19-Jun-07
2403    ---------------------
2404    
2405     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2406        which is apparently normally available under Windows.
2407    
2408     2. Re-jig the pcregrep tests with different newline settings in an attempt
2409        to make them independent of the local environment's newline setting.
2410    
2411     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2412    
2413     4. Some of the "internals" tests were previously cut out when the link size
2414        was not 2, because the output contained actual offsets. The recent new
2415        "Z" feature of pcretest means that these can be cut out, making the tests
2416        usable with all link sizes.
2417    
2418     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2419        stack recursion. This gives a massive performance boost under BSD, but just
2420        a small improvement under Linux. However, it saves one field in the frame
2421        in all cases.
2422    
2423     6. Added more features from the forthcoming Perl 5.10:
2424    
2425        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2426            recursion call. It refers to the nth most recently opened parentheses.
2427    
2428        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2429            to be opened parentheses.
2430    
2431        (c) Conditions that refer to capturing parentheses can be specified
2432            relatively, for example, (?(-2)... or (?(+3)...
2433    
2434        (d) \K resets the start of the current match so that everything before
2435            is not part of it.
2436    
2437        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2438    
2439        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2440            reference syntax.
2441    
2442        (g) (?| introduces a group in which the numbering of parentheses in each
2443            alternative starts with the same number.
2444    
2445        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2446    
2447     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2448        PCRE_INFO_JCHANGED.
2449    
2450     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2451        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2452        for detecting groups that can match an empty string.
2453    
2454     9. A pattern with a very large number of alternatives (more than several
2455        hundred) was running out of internal workspace during the pre-compile
2456        phase, where pcre_compile() figures out how much memory will be needed. A
2457        bit of new cunning has reduced the workspace needed for groups with
2458        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2459        workspace instead of running out of the 4096 that are available.
2460    
2461    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2462    
2463    11. Applied patch from Google to remove an optimization that didn't quite work.
2464        The report of the bug said:
2465    
2466          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2467          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2468          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2469    
2470    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2471        it matched the wrong number of bytes.
2472    
2473    
2474    Version 7.1 24-Apr-07
2475    ---------------------
2476    
2477     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2478        that is more "standard", making use of automake and other Autotools. There
2479        is some re-arrangement of the files and adjustment of comments consequent
2480        on this.
2481    
2482     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2483        for recursive directory scanning broke on some systems because the files
2484        are not scanned in any specific order and on different systems the order
2485        was different. A call to "sort" has been inserted into RunGrepTest for the
2486        approprate test as a short-term fix. In the longer term there may be an
2487        alternative.
2488    
2489     3. I had an email from Eric Raymond about problems translating some of PCRE's
2490        man pages to HTML (despite the fact that I distribute HTML pages, some
2491        people do their own conversions for various reasons). The problems
2492        concerned the use of low-level troff macros .br and .in. I have therefore
2493        removed all such uses from the man pages (some were redundant, some could
2494        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2495        HTML has been updated to handle .nf/.fi and to complain if it encounters
2496        .br or .in.
2497    
2498     4. Updated comments in configure.ac that get placed in config.h.in and also
2499        arranged for config.h to be included in the distribution, with the name
2500        config.h.generic, for the benefit of those who have to compile without
2501        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2502    
2503     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2504        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2505        makevp.bat for the current PCRE, using the additional files
2506        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2507    
2508     6. A Windows user reported a minor discrepancy with test 2, which turned out
2509        to be caused by a trailing space on an input line that had got lost in his
2510        copy. The trailing space was an accident, so I've just removed it.
2511    
2512     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2513        that is needed.
2514    
2515     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2516        as "const" (a) because they are and (b) because it helps the PHP
2517        maintainers who have recently made a script to detect big data structures
2518        in the php code that should be moved to the .rodata section. I remembered
2519        to update Builducptable as well, so it won't revert if ucptable.h is ever
2520        re-created.
2521    
2522     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2523        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2524        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2525        support is not required. This saves 1.5-2K of code, which is important in
2526        some applications.
2527    
2528        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2529        so as not to refer to the tables, even though these functions will never be
2530        called when UTF-8 support is disabled. Otherwise there are problems with a
2531        shared library.
2532    
2533    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2534    
2535        (a) It was defining its arguments as char * instead of void *.
2536    
2537        (b) It was assuming that all moves were upwards in memory; this was true
2538            a long time ago when I wrote it, but is no longer the case.
2539    
2540        The emulated memove() is provided for those environments that have neither
2541        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2542        is clearly not the case, as these two bugs were recently reported.
2543    
2544    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2545        and Detrail to create the HTML documentation, the .txt form of the man
2546        pages, and it removes trailing spaces from listed files. It also creates
2547        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2548        case, it wraps all the #defines with #ifndefs. This script should be run
2549        before "make dist".
2550    
2551    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2552        with Unicode property support.
2553    
2554        (a) For a maximizing quantifier, if the two different cases of the
2555            character were of different lengths in their UTF-8 codings (there are
2556            some cases like this - I found 11), and the matching function had to
2557            back up over a mixture of the two cases, it incorrectly assumed they
2558            were both the same length.
2559    
2560        (b) When PCRE was configured to use the heap rather than the stack for
2561            recursion during matching, it was not correctly preserving the data for
2562            the other case of a UTF-8 character when checking ahead for a match
2563            while processing a minimizing repeat. If the check also involved
2564            matching a wide character, but failed, corruption could cause an
2565            erroneous result when trying to check for a repeat of the original
2566            character.
2567    
2568    13. Some tidying changes to the testing mechanism:
2569    
2570        (a) The RunTest script now detects the internal link size and whether there
2571            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2572            values substituted by "configure". (The RunGrepTest script already did
2573            this for UTF-8.) The configure.ac script no longer substitutes the
2574            relevant variables.
2575    
2576        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2577            with length and offset values. This means that the output is different
2578            for different internal link sizes. Test 2 is skipped for link sizes
2579            other than 2 because of this, bypassing the problem. Unfortunately,
2580            there was also a test in test 3 (the locale tests) that used /B and
2581            failed for link sizes other than 2. Rather than cut the whole test out,
2582            I have added a new /Z option to pcretest that replaces the length and
2583            offset values with spaces. This is now used to make test 3 independent
2584            of link size. (Test 2 will be tidied up later.)
2585    
2586    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2587        segmentation fault instead of returning the appropriate error message.
2588    
2589    15. In multiline mode when the newline sequence was set to "any", the pattern
2590        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2591        This doesn't seem right; it now treats the CRLF combination as the line
2592        ending, and so does not match in that case. It's only a pattern such as ^$
2593        that would hit this one: something like ^ABC$ would have failed after \r
2594        and then tried again after \r\n.
2595    
2596    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2597        in an attempt to make files that differ only in their line terminators
2598        compare equal. This works on Linux.
2599    
2600    17. Under certain error circumstances pcregrep might try to free random memory
2601        as it exited. This is now fixed, thanks to valgrind.
2602    
2603    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2604        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2605        was because its rules for how to advance for /g after matching an empty
2606        string at the end of a line did not allow for this case. They now check for
2607        it specially.
2608    
2609    20. pcretest is supposed to handle patterns and data of any length, by
2610        extending its buffers when necessary. It was getting this wrong when the
2611        buffer for a data line had to be extended.
2612    
2613    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2614        CRLF as a newline sequence.
2615    
2616    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2617        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2618        I have nevertheless tidied it up.
2619    
2620    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2621    
2622    24. Added a man page for pcre-config.
2623    
2624    
2625    Version 7.0 19-Dec-06
2626    ---------------------
2627    
2628     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2629        moving to gcc 4.1.1.
2630    
2631     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2632        sys/time.h, which is documented as needed for this function. It doesn't
2633        seem to matter on Linux, but it showed up on some releases of OS X.
2634    
2635     3. It seems that there are systems where bytes whose values are greater than
2636        127 match isprint() in the "C" locale. The "C" locale should be the
2637        default when a C program starts up. In most systems, only ASCII printing
2638        characters match isprint(). This difference caused the output from pcretest
2639        to vary, making some of the tests fail. I have changed pcretest so that:
2640    
2641        (a) When it is outputting text in the compiled version of a pattern, bytes
2642            other than 32-126 are always shown as hex escapes.
2643    
2644        (b) When it is outputting text that is a matched part of a subject string,
2645            it does the same, unless a different locale has been set for the match
2646            (using the /L modifier). In this case, it uses isprint() to decide.
2647    
2648     4. Fixed a major bug that caused incorrect computation of the amount of memory
2649        required for a compiled pattern when options that changed within the
2650        pattern affected the logic of the preliminary scan that determines the
2651        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2652        that the computed length was too small. The symptoms of this bug were
2653        either the PCRE error "internal error: code overflow" from pcre_compile(),
2654        or a glibc crash with a message such as "pcretest: free(): invalid next
2655        size (fast)". Examples of patterns that provoked this bug (shown in
2656        pcretest format) are:
2657    
2658          /(?-x: )/x
2659          /(?x)(?-x: \s*#\s*)/
2660          /((?i)[\x{c0}])/8
2661          /(?i:[\x{c0}])/8
2662    
2663        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2664        is now done differently.
2665    
2666     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2667        wrapper classes; (b) implement a new function in the C++ scanner that is
2668        more efficient than the old way of doing things because it avoids levels of
2669        recursion in the regex matching; (c) add a paragraph to the documentation
2670        for the FullMatch() function.
2671    
2672     6. The escape sequence \n was being treated as whatever was defined as
2673        "newline". Not only was this contrary to the documentation, which states
2674        that \n is character 10 (hex 0A), but it also went horribly wrong when
2675        "newline" was defined as CRLF. This has been fixed.
2676    
2677     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2678        was being set to -1 for the "end of line" case (supposedly a value that no
2679        character can have). Though this value is never used (the check for end of
2680        line is "zero bytes in current character"), it caused compiler complaints.
2681        I've changed it to 0xffffffff.
2682    
2683     8. In pcre_version.c, the version string was being built by a sequence of
2684        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2685        string (as it is for production releases) called a macro with an empty
2686        argument. The C standard says the result of this is undefined. The gcc
2687        compiler treats it as an empty string (which was what was wanted) but it is
2688        reported that Visual C gives an error. The source has been hacked around to
2689        avoid this problem.
2690    
2691     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2692        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2693        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2694        of them did).
2695    
2696    10. Originally, pcretest opened its input and output without "b"; then I was
2697        told that "b" was needed in some environments, so it was added for release
2698        5.0 to both the input and output. (It makes no difference on Unix-like
2699        systems.) Later I was told that it is wrong for the input on Windows. I've
2700        now abstracted the modes into two macros, to make it easier to fiddle with
2701        them, and removed "b" from the input mode under Windows.
2702    
2703    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2704    
2705    12. Added -help and --help to pcretest as an official way of being reminded
2706        of the options.
2707    
2708    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2709        and pcrecpp.cc because they annoy compilers at high warning levels.
2710    
2711    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2712    
2713    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2714        scripts are not C programs :-) and which was not noticed because it works
2715        on Linux.
2716    
2717    16. pcretest is supposed to handle any length of pattern and data line (as one
2718        line or as a continued sequence of lines) by extending its input buffer if
2719        necessary. This feature was broken for very long pattern lines, leading to
2720        a string of junk being passed to pcre_compile() if the pattern was longer
2721        than about 50K.
2722    
2723    17. I have done a major re-factoring of the way pcre_compile() computes the
2724        amount of memory needed for a compiled pattern. Previously, there was code
2725        that made a preliminary scan of the pattern in order to do this. That was
2726        OK when PCRE was new, but as the facilities have expanded, it has become
2727        harder and harder to keep it in step with the real compile phase, and there
2728        have been a number of bugs (see for example, 4 above). I have now found a
2729        cunning way of running the real compile function in a "fake" mode that
2730        enables it to compute how much memory it would need, while actually only
2731        ever using a few hundred bytes of working memory and without too many
2732        tests of the mode. This should make future maintenance and development
2733        easier. A side effect of this work is that the limit of 200 on the nesting
2734        depth of parentheses has been removed (though this was never a serious
2735        limitation, I suspect). However, there is a downside: pcre_compile() now
2736        runs more slowly than before (30% or more, depending on the pattern). I
2737        hope this isn't a big issue. There is no effect on runtime performance.
2738    
2739    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2740        newline (only possible for the last line of a file) and it was a
2741        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2742    
2743    19. Added additional timing features to pcretest. (1) The -tm option now times
2744        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2745        separate command line item, by a number that specifies the number of
2746        repeats to use when timing. The default is 50000; this gives better
2747        precision, but takes uncomfortably long for very large patterns.
2748    
2749    20. Extended pcre_study() to be more clever in cases where a branch of a
2750        subpattern has no definite first character. For example, (a*|b*)[cd] would
2751        previously give no result from pcre_study(). Now it recognizes that the
2752        first character must be a, b, c, or d.
2753    
2754    21. There was an incorrect error "recursive call could loop indefinitely" if
2755        a subpattern (or the entire pattern) that was being tested for matching an
2756        empty string contained only one non-empty item after a nested subpattern.
2757        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2758        incorrectly, because the \d was being skipped in the check.
2759    
2760    22. The pcretest program now has a new pattern option /B and a command line
2761        option -b, which is equivalent to adding /B to every pattern. This causes
2762        it to show the compiled bytecode, without the additional information that
2763        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2764        is the same as /B/I).
2765    
2766    23. A new optimization is now able automatically to treat some sequences such
2767        as a*b as a*+b. More specifically, if something simple (such as a character
2768        or a simple class like \d) has an unlimited quantifier, and is followed by
2769        something that cannot possibly match the quantified thing, the quantifier
2770        is automatically "possessified".
2771    
2772    24. A recursive reference to a subpattern whose number was greater than 39
2773        went wrong under certain circumstances in UTF-8 mode. This bug could also
2774        have affected the operation of pcre_study().
2775    
2776    25. Realized that a little bit of performance could be had by replacing
2777        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2778    
2779    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2780    
2781    27. Possessive quantifiers such as a++ were previously implemented by turning
2782        them into atomic groups such as ($>a+). Now they have their own opcodes,
2783        which improves performance. This includes the automatically created ones
2784        from 23 above.
2785    
2786    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2787        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2788        the first matched character to be a colon. This applied both to named and
2789        numbered groups.
2790    
2791    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2792    
2793    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2794        building PCRE on Windows easier, so I have included it in the distribution.
2795    
2796    31. There is now a check in pcretest against a ridiculously large number being
2797        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2798        loop, the loop is abandoned.
2799    
2800    32. Forward references to subpatterns in conditions such as (?(2)...) where
2801        subpattern 2 is defined later cause pcre_compile() to search forwards in
2802        the pattern for the relevant set of parentheses. This search went wrong
2803        when there were unescaped parentheses in a character class, parentheses
2804        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2805    
2806    33. "Subroutine" calls and backreferences were previously restricted to
2807        referencing subpatterns earlier in the regex. This restriction has now
2808        been removed.
2809    
2810    34. Added a number of extra features that are going to be in Perl 5.10. On the
2811        whole, these are just syntactic alternatives for features that PCRE had
2812        previously implemented using the Python syntax or my own invention. The
2813        other formats are all retained for compatibility.
2814    
2815        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2816            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2817            also .NET compatible.
2818    
2819        (b) A recursion or subroutine call to a named group can now be defined as
2820            (?&name) as well as (?P>name).
2821    
2822        (c) A backreference to a named group can now be defined as \k<name> or
2823            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2824            5.10, are also .NET compatible.
2825    
2826        (d) A conditional reference to a named group can now use the syntax
2827            (?(<name>) or (?('name') as well as (?(name).
2828    
2829        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2830            groups (named and numbered) that are never evaluated inline, but can be
2831            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2832            is always false. There may be only one alternative in such a group.
2833    
2834        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2835            as the simple (?(R). The condition is true only if the most recent
2836            recursion is that of the given number or name. It does not search out
2837            through the entire recursion stack.
2838    
2839        (g) The escape \gN or \g{N} has been added, where N is a positive or
2840            negative number, specifying an absolute or relative reference.
2841    
2842    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2843        some "unreachable code" warnings.
2844    
2845    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2846        things, this adds five new scripts.
2847    
2848    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2849        There were also incompatibilities regarding the handling of \Q..\E inside
2850        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2851        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2852    
2853    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2854        matches an empty string, and forcibly breaks the loop. There were bugs in
2855        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2856        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2857        separate and independent bugs (that affected different cases) have been
2858        fixed.
2859    
2860    39. Refactored the code to abolish the use of different opcodes for small
2861        capturing bracket numbers. This is a tidy that I avoided doing when I
2862        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2863        The new approach is not only tidier, it makes it possible to reduce the
2864        memory needed to fix the previous bug (38).
2865    
2866    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2867        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2868        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2869        mode.
2870    
2871    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2872        report.
2873    
2874    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2875        copy construction and assignment in the C++ wrapper.
2876    
2877    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2878        couple of bugs that could have given wrong results in the "--newline=crlf"
2879        case.
2880    
2881    44. Added a number of casts and did some reorganization of signed/unsigned int
2882        variables following suggestions from Dair Grant. Also renamed the variable
2883        "this" as "item" because it is a C++ keyword.
2884    
2885    45. Arranged for dftables to add
2886    
2887          #include "pcre_internal.h"
2888    
2889        to pcre_chartables.c because without it, gcc 4.x may remove the array
2890        definition from the final binary if PCRE is built into a static library and
2891        dead code stripping is activated.
2892    
2893    46. For an unanchored pattern, if a match attempt fails at the start of a
2894        newline sequence, and the newline setting is CRLF or ANY, and the next two
2895        characters are CRLF, advance by two characters instead of one.
2896    
2897    
2898    Version 6.7 04-Jul-06
2899    ---------------------
2900    
2901     1. In order to handle tests when input lines are enormously long, pcretest has
2902        been re-factored so that it automatically extends its buffers when
2903        necessary. The code is crude, but this _is_ just a test program. The
2904        default size has been increased from 32K to 50K.
2905    
2906     2. The code in pcre_study() was using the value of the re argument before
2907        testing it for NULL. (Of course, in any sensible call of the function, it
2908        won't be NULL.)
2909    
2910     3. The memmove() emulation function in pcre_internal.h, which is used on
2911        systems that lack both memmove() and bcopy() - that is, hardly ever -
2912        was missing a "static" storage class specifier.
2913    
2914     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2915        containing an extended class (one that cannot be represented by a bitmap
2916        because it contains high-valued characters or Unicode property items, e.g.
2917        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2918        pattern, but PCRE should not loop if you do not (it no longer does).
2919        [Detail: two cases were found: (a) a repeated subpattern containing an
2920        extended class; (b) a recursive reference to a subpattern that followed a
2921        previous extended class. It wasn't skipping over the extended class
2922        correctly when UTF-8 mode was not set.]
2923    
2924     5. A negated single-character class was not being recognized as fixed-length
2925        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2926        compile error "lookbehind assertion is not fixed length".
2927    
2928     6. The RunPerlTest auxiliary script was showing an unexpected difference
2929        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2930        write a Perl script that can interpret lines of an input file either as
2931        byte characters or as UTF-8, which is what "perltest" was being required to
2932        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2933        can't do is switch easily at run time between having the "use utf8;" pragma
2934        or not. In the end, I fudged it by using the RunPerlTest script to insert
2935        "use utf8;" explicitly for the UTF-8 tests.
2936    
2937     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2938        the end of the subject string, contrary to the documentation and to what
2939        Perl does. This was true of both matching functions. Now it matches only at
2940        the start of the subject and immediately after *internal* newlines.
2941    
2942     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2943        a pointer to an int instead of a pointer to an unsigned long int. This
2944        caused problems on 64-bit systems.
2945    
2946     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2947        instance of the 'standard' template library not being so standard".
2948    
2949    10. There was no check on the number of named subpatterns nor the maximum
2950        length of a subpattern name. The product of these values is used to compute
2951        the size of the memory block for a compiled pattern. By supplying a very
2952        long subpattern name and a large number of named subpatterns, the size
2953        computation could be caused to overflow. This is now prevented by limiting
2954        the length of names to 32 characters, and the number of named subpatterns
2955        to 10,000.
2956    
2957    11. Subpatterns that are repeated with specific counts have to be replicated in
2958        the compiled pattern. The size of memory for this was computed from the
2959        length of the subpattern and the repeat count. The latter is limited to
2960        65535, but there was no limit on the former, meaning that integer overflow
2961        could in principle occur. The compiled length of a repeated subpattern is
2962        now limited to 30,000 bytes in order to prevent this.
2963    
2964    12. Added the optional facility to have named substrings with the same name.
2965    
2966    13. Added the ability to use a named substring as a condition, using the
2967        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2968        are numbers (not recommended). Forward references are permitted.
2969    
2970    14. Added forward references in named backreferences (if you see what I mean).
2971    
2972    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2973        pattern could run off the end of the subject. For example, the pattern
2974        "(?s)(.{1,5})"8 did this with the subject "ab".
2975    
2976    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2977        PCRE_CASELESS was set when matching characters that were quantified with ?
2978        or *.
2979    
2980    17. A character class other than a single negated character that had a minimum
2981        but no maximum quantifier - for example [ab]{6,} - was not handled
2982        correctly by pce_dfa_exec(). It would match only one character.
2983    
2984    18. A valid (though odd) pattern that looked like a POSIX character
2985        class but used an invalid character after [ (for example [[,abc,]]) caused
2986        pcre_compile() to give the error "Failed: internal error: code overflow" or
2987        in some cases to crash with a glibc free() error. This could even happen if
2988        the pattern terminated after [[ but there just happened to be a sequence of
2989        letters, a binary zero, and a closing ] in the memory that followed.
2990    
2991    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2992        over the years. Originally (before any Unicode support), just the bottom 8
2993        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2994        output from "man perlunicode" includes this:
2995    
2996          The regular expression compiler produces polymorphic opcodes.  That
2997          is, the pattern adapts to the data and automatically switches to
2998          the Unicode character scheme when presented with Unicode data--or
2999          instead uses a traditional byte scheme when presented with byte
3000          data.
3001    
3002        Sadly, a wide octal escape does not cause a switch, and in a string with
3003        no other multibyte characters, these octal escapes are treated as before.
3004        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
3005        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3006        Unicode string.
3007    
3008        I have not perpetrated such confusion in PCRE. Up till now, it took just
3009        the bottom 8 bits, as in old Perl. I have now made octal escapes with
3010        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3011        translate to the appropriate multibyte character.
3012    
3013    29. Applied some refactoring to reduce the number of warnings from Microsoft
3014        and Borland compilers. This has included removing the fudge introduced
3015        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3016        a warning about an unused variable.
3017    
3018    21. PCRE has not included VT (character 0x0b) in the set of whitespace
3019        characters since release 4.0, because Perl (from release 5.004) does not.
3020        [Or at least, is documented not to: some releases seem to be in conflict
3021        with the documentation.] However, when a pattern was studied with
3022        pcre_study() and all its branches started with \s, PCRE still included VT
3023        as a possible starting character. Of course, this did no harm; it just
3024        caused an unnecessary match attempt.
3025    
3026    22. Removed a now-redundant internal flag bit that recorded the fact that case
3027        dependency changed within the pattern. This was once needed for "required
3028        byte" processing, but is no longer used. This recovers a now-scarce options
3029        bit. Also moved the least significant internal flag bit to the most-
3030        significant bit of the word, which was not previously used (hangover from
3031        the days when it was an int rather than a uint) to free up another bit for
3032        the future.
3033    
3034    23. Added support for CRLF line endings as well as CR and LF. As well as the
3035        default being selectable at build time, it can now be changed at runtime
3036        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3037        specify that it is scanning data with non-default line endings.
3038    
3039    24. Changed the definition of CXXLINK to make it agree with the definition of
3040        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3041    
3042    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3043        recursions. This makes a big different to stack usage for some patterns.
3044    
3045    26. If a subpattern containing a named recursion or subroutine reference such
3046        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3047        the space required for the compiled pattern went wrong and gave too small a
3048        value. Depending on the environment, this could lead to "Failed: internal
3049        error: code overflow at offset 49" or "glibc detected double free or
3050        corruption" errors.
3051    
3052    27. Applied patches from Google (a) to support the new newline modes and (b) to
3053        advance over multibyte UTF-8 characters in GlobalReplace.
3054    
3055    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3056        difference for some implementation of PCRE in some Windows version.
3057    
3058    29. Added some extra testing facilities to pcretest:
3059    
3060        \q<number>   in a data line sets the "match limit" value
3061        \Q<number>   in a data line sets the "match recursion limt" value
3062        -S <number>  sets the stack size, where <number> is in megabytes
3063    
3064        The -S option isn't available for Windows.
3065    
3066    
3067    Version 6.6 06-Feb-06
3068    ---------------------
3069    
3070     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3071        in pcreposix.h. I have copied the definition from pcre.h.
3072    
3073     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3074        because pcre.h is no longer a built file.
3075    
3076     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3077        not normally included in the compiled code.
3078    
3079    
3080    Version 6.5 01-Feb-06
3081    ---------------------
3082    
3083     1. When using the partial match feature with pcre_dfa_exec(), it was not
3084        anchoring the second and subsequent partial matches at the new starting
3085        point. This could lead to incorrect results. For example, with the pattern
3086        /1234/, partially matching against "123" and then "a4" gave a match.
3087    
3088     2. Changes to pcregrep:
3089    
3090        (a) All non-match returns from pcre_exec() were being treated as failures
3091            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3092            error message is output. Some extra information is given for the
3093            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3094            probably the only errors that are likely to be caused by users (by
3095            specifying a regex that has nested indefinite repeats, for instance).
3096            If there are more than 20 of these errors, pcregrep is abandoned.
3097    
3098        (b) A binary zero was treated as data while matching, but terminated the
3099            output line if it was written out. This has been fixed: binary zeroes
3100            are now no different to any other data bytes.
3101    
3102        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3103            used to set a locale for matching. The --locale=xxxx long option has
3104            been added (no short equivalent) to specify a locale explicitly on the
3105            pcregrep command, overriding the environment variables.
3106    
3107        (d) When -B was used with -n, some line numbers in the output were one less
3108            than they should have been.
3109    
3110        (e) Added the -o (--only-matching) option.
3111    
3112        (f) If -A or -C was used with -c (count only), some lines of context were
3113            accidentally printed for the final match.
3114    
3115        (g) Added the -H (--with-filename) option.
3116    
3117        (h) The combination of options -rh failed to suppress file names for files
3118            that were found from directory arguments.
3119    
3120        (i) Added the -D (--devices) and -d (--directories) options.
3121    
3122        (j) Added the -F (--fixed-strings) option.
3123    
3124        (k) Allow "-" to be used as a file name for -f as well as for a data file.
3125    
3126        (l) Added the --colo(u)r option.
3127    
3128        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3129            is not present by default.
3130    
3131     3. A nasty bug was discovered in the handling of recursive patterns, that is,
3132        items such as (?R) or (?1), when the recursion could match a number of
3133        alternatives. If it matched one of the alternatives, but subsequently,
3134        outside the recursion, there was a failure, the code tried to back up into
3135        the recursion. However, because of the way PCRE is implemented, this is not
3136        possible, and the result was an incorrect result from the match.
3137    
3138        In order to prevent this happening, the specification of recursion has
3139        been changed so that all such subpatterns are automatically treated as
3140        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3141    
3142     4. I had overlooked the fact that, in some locales, there are characters for
3143        which isalpha() is true but neither isupper() nor islower() are true. In
3144        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3145        and ordfeminine) are like this. This affected the treatment of \w and \W
3146        when they appeared in character classes, but not when they appeared outside
3147        a character class. The bit map for "word" characters is now created
3148        separately from the results of isalnum() instead of just taking it from the
3149        upper, lower, and digit maps. (Plus the underscore character, of course.)
3150    
3151     5. The above bug also affected the handling of POSIX character classes such as
3152        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3153        permanent tables. Instead, the bit maps for such a class were previously
3154        created as the appropriate unions of the upper, lower, and digit bitmaps.
3155        Now they are created by subtraction from the [[:word:]] class, which has
3156        its own bitmap.
3157    
3158     6. The [[:blank:]] character class matches horizontal, but not vertical space.
3159        It is created by subtracting the vertical space characters (\x09, \x0a,
3160        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3161        subtraction was done in the overall bitmap for a character class, meaning
3162        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3163        be recognized. This bug has been fixed.
3164    
3165     7. Patches from the folks at Google:
3166    
3167          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3168          real life, but is still worth protecting against".
3169    
3170          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3171          regular expressions".
3172    
3173          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3174          have it.
3175    
3176          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3177          "configure" and the latter not, in order to fix a problem somebody had
3178          with compiling the Arg class on HP-UX.
3179    
3180          (e) Improve the error-handling of the C++ wrapper a little bit.
3181    
3182          (f) New tests for checking recursion limiting.
3183    
3184     8. The pcre_memmove() function, which is used only if the environment does not
3185        have a standard memmove() function (and is therefore rarely compiled),
3186        contained two bugs: (a) use of int instead of size_t, and (b) it was not
3187        returning a result (though PCRE never actually uses the result).
3188    
3189     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3190        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3191        returned instead of calling malloc() with an overflowing number that would
3192        most likely cause subsequent chaos.
3193    
3194    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3195    
3196    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3197        with this option is matched, the nmatch and pmatch options of regexec() are
3198        ignored.
3199    
3200    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3201        provided in case anyone wants to the the POSIX interface with UTF-8
3202        strings.
3203    
3204    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3205        C++ linking (needed for some HP-UX environments).
3206    
3207    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3208        (unused parameter) and in the pcre_printint() function (omitted "default"
3209        switch label when the default is to do nothing).
3210    
3211    15. Added some code to make it possible, when PCRE is compiled as a C++
3212        library, to replace subject pointers for pcre_exec() with a smart pointer
3213        class, thus making it possible to process discontinuous strings.
3214    
3215    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3216        much the same function. They were added by different people who were trying
3217        to make PCRE easy to compile on non-Unix systems. It has been suggested
3218        that PCRE_EXPORT be abolished now that there is more automatic apparatus
3219        for compiling on Windows systems. I have therefore replaced it with
3220        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3221        defaults to "extern" for C or "extern C" for C++, which works fine on
3222        Unix-like systems. It is now possible to override the value of PCRE_DATA_
3223        SCOPE with something explicit in config.h. In addition:
3224    
3225        (a) pcreposix.h still had just "extern" instead of either of these macros;
3226            I have replaced it with PCRE_DATA_SCOPE.
3227    
3228        (b) Functions such as _pcre_xclass(), which are internal to the library,
3229            but external in the C sense, all had PCRE_EXPORT in their definitions.
3230            This is apparently wrong for the Windows case, so I have removed it.
3231            (It makes no difference on Unix-like systems.)
3232    
3233    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3234        of recursive calls to match(). This is different to MATCH_LIMIT because
3235        that limits the total number of calls to match(), not all of which increase
3236        the depth of recursion. Limiting the recursion depth limits the amount of
3237        stack (or heap if NO_RECURSE is set) that is used. The default can be set
3238        when PCRE is compiled, and changed at run time. A patch from Google adds
3239        this functionality to the C++ interface.
3240    
3241    18. Changes to the handling of Unicode character properties:
3242    
3243        (a) Updated the table to Unicode 4.1.0.
3244    
3245        (b) Recognize characters that are not in the table as "Cn" (undefined).
3246    
3247        (c) I revised the way the table is implemented to a much improved format
3248            which includes recognition of ranges. It now supports the ranges that
3249            are defined in UnicodeData.txt, and it also amalgamates other
3250            characters into ranges. This has reduced the number of entries in the
3251            table from around 16,000 to around 3,000, thus reducing its size
3252            considerably. I realized I did not need to use a tree structure after
3253            all - a binary chop search is just as efficient. Having reduced the
3254            number of entries, I extended their size from 6 bytes to 8 bytes to
3255            allow for more data.
3256    
3257        (d) Added support for Unicode script names via properties such as \p{Han}.
3258    
3259    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3260        matching that character.
3261    
3262    20. When matching a repeated Unicode property with a minimum greater than zero,
3263        (for example \pL{2,}), PCRE could look past the end of the subject if it
3264        reached it while seeking the minimum number of characters. This could
3265        happen only if some of the characters were more than one byte long, because
3266        there is a check for at least the minimum number of bytes.
3267    
3268    21. Refactored the implementation of \p and \P so as to be more general, to
3269        allow for more different types of property in future. This has changed the
3270        compiled form incompatibly. Anybody with saved compiled patterns that use
3271        \p or \P will have to recompile them.
3272    
3273    22. Added "Any" and "L&" to the supported property types.
3274    
3275    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3276        but give a compile time error if the value is greater than 0xff.
3277    
3278    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3279        accidentally not being installed or uninstalled.
3280    
3281    25. The pcre.h file was built from pcre.h.in, but the only changes that were
3282        made were to insert the current release number. This seemed silly, because
3283        it made things harder for people building PCRE on systems that don't run
3284        "configure". I have turned pcre.h into a distributed file, no longer built
3285        by "configure", with the version identification directly included. There is
3286        no longer a pcre.h.in file.
3287    
3288        However, this change necessitated a change to the pcre-config script as
3289        well. It is built from pcre-config.in, and one of the substitutions was the
3290        release number. I have updated configure.ac so that ./configure now finds
3291        the release number by grepping pcre.h.
3292    
3293    26. Added the ability to run the tests under valgrind.
3294    
3295    
3296    Version 6.4 05-Sep-05
3297    ---------------------
3298    
3299     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3300        "--" to be printed when multiple files were scanned, even when none of the
3301        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3302        consider it to be a bug, and have restored the previous behaviour.
3303    
3304     2. A couple of code tidies to get rid of compiler warnings.
3305    
3306     3. The pcretest program used to cheat by referring to symbols in the library
3307        whose names begin with _pcre_. These are internal symbols that are not
3308        really supposed to be visible externally, and in some environments it is
3309        possible to suppress them. The cheating is now confined to including
3310        certain files from the library's source, which is a bit cleaner.
3311    
3312     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3313        file's purpose clearer.
3314    
3315     5. Reorganized pcre_ucp_findchar().
3316    
3317    
3318    Version 6.3 15-Aug-05
3319    ---------------------
3320    
3321     1. The file libpcre.pc.in did not have general read permission in the tarball.
3322    
3323     2. There were some problems when building without C++ support:
3324    
3325        (a) If C++ support was not built, "make install" and "make test" still
3326            tried to test it.
3327    
3328        (b) There were problems when the value of CXX was explicitly set. Some
3329            changes have been made to try to fix these, and ...
3330    
3331        (c) --disable-cpp can now be used to explicitly disable C++ support.
3332    
3333        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3334            backslash in a target when C++ was disabled. This confuses some
3335            versions of "make", apparently. Using an intermediate variable solves
3336            this. (Same for CPP_LOBJ.)
3337    
3338     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3339        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3340        necessary on certain architectures.
3341    
3342     4. Added a setting of -export-symbols-regex to the link command to remove
3343        those symbols that are exported in the C sense, but actually are local
3344        within the library, and not documented. Their names all begin with
3345        "_pcre_". This is not a perfect job, because (a) we have to except some
3346        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3347        available (and never for static libraries). I have made a note to try to
3348        find a way round (a) in the future.
3349    
3350    
3351    Version 6.2 01-Aug-05
3352    ---------------------
3353    
3354     1. There was no test for integer overflow of quantifier values. A construction
3355        such as {1111111111111111} would give undefined results. What is worse, if
3356        a minimum quantifier for a parenthesized subpattern overflowed and became
3357        negative, the calculation of the memory size went wrong. This could have
3358        led to memory overwriting.
3359    
3360     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3361    
3362     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3363        operating environments where this matters.
3364    
3365     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3366        PCRE options from within the C++ wrapper.
3367    
3368     5. Named capturing subpatterns were not being correctly counted when a pattern
3369        was compiled. This caused two problems: (a) If there were more than 100
3370        such subpatterns, the calculation of the memory needed for the whole
3371        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3372        back references of the form \12, where the number was greater than 9, were
3373        not recognized as back references, even though there were sufficient
3374        previous subpatterns.
3375    
3376     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3377        versions of gcc, e.g. 2.95.4.
3378    
3379    
3380    Version 6.1 21-Jun-05
3381    ---------------------
3382    
3383     1. There was one reference to the variable "posix" in pcretest.c that was not
3384        surrounded by "#if !defined NOPOSIX".
3385    
3386     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3387        the cross-check on the old pcre_info() function, for the benefit of the
3388        cut-down version of PCRE that is currently imported into Exim.
3389    
3390     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3391        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3392        patterns that start (?i) but I don't think that matters. The use of (?i) is
3393        just an example; this all applies to the other options as well.
3394    
3395     4. Since libtool seems to echo the compile commands it is issuing, the output
3396        from "make" can be reduced a bit by putting "@" in front of each libtool
3397        compile command.
3398    
3399     5. Patch from the folks at Google for configure.in to be a bit more thorough
3400        in checking for a suitable C++ installation before trying to compile the
3401        C++ stuff. This should fix a reported problem when a compiler was present,
3402        but no suitable headers.
3403    
3404     6. The man pages all had just "PCRE" as their title. I have changed them to
3405        be the relevant file name. I have also arranged that these names are
3406        retained in the file doc/pcre.txt, which is a concatenation in text format
3407        of all the man pages except the little individual ones for each function.
3408    
3409     7. The NON-UNIX-USE file had not been updated for the different set of source
3410        files that come with release 6. I also added a few comments about the C++
3411        wrapper.
3412    
3413    
3414    Version 6.0 07-Jun-05
3415    ---------------------
3416    
3417     1. Some minor internal re-organization to help with my DFA experiments.
3418    
3419     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3420        didn't matter for the library itself when fully configured, but did matter
3421        when compiling without UCP support, or within Exim, where the ucp files are
3422        not imported.
3423    
3424     3. Refactoring of the library code to split up the various functions into
3425        different source modules. The addition of the new DFA matching code (see
3426        below) to a single monolithic source would have made it really too
3427        unwieldy, quite apart from causing all the code to be include in a
3428        statically linked application, when only some functions are used. This is
3429        relevant even without the DFA addition now that patterns can be compiled in
3430        one application and matched in another.
3431    
3432        The downside of splitting up is that there have to be some external
3433        functions and data tables that are used internally in different modules of
3434        the library but which are not part of the API. These have all had their
3435        names changed to start with "_pcre_" so that they are unlikely to clash
3436        with other external names.
3437    
3438     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3439        a different (DFA) algorithm. Although it is slower than the original
3440        function, it does have some advantages for certain types of matching
3441        problem.
3442    
3443     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3444        including restarting after a partial match.
3445    
3446     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3447        defined when compiling for Windows was sent to me. I have put it into the
3448        code, though I have no means of testing or verifying it.
3449    
3450     7. Added the pcre_refcount() auxiliary function.
3451    
3452     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3453        match before or at the first newline in the subject string. In pcretest,
3454        the /f option on a pattern can be used to set this.
3455    
3456     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3457        would behave wrongly. This has been present in PCRE since release 4.0.
3458    
3459    10. A number of changes to the pcregrep command:
3460    
3461        (a) Refactored how -x works; insert ^(...)$ instead of setting
3462            PCRE_ANCHORED and checking the length, in preparation for adding
3463            something similar for -w.
3464    
3465        (b) Added the -w (match as a word) option.
3466    
3467        (c) Refactored the way lines are read and buffered so as to have more
3468            than one at a time available.
3469    
3470        (d) Implemented a pcregrep test script.
3471    
3472        (e) Added the -M (multiline match) option. This allows patterns to match
3473            over several lines of the subject. The buffering ensures that at least
3474            8K, or the rest of the document (whichever is the shorter) is available
3475            for matching (and similarly the previous 8K for lookbehind assertions).
3476    
3477        (f) Changed the --help output so that it now says
3478    
3479              -w, --word-regex(p)
3480    
3481            instead of two lines, one with "regex" and the other with "regexp"
3482            because that confused at least one person since the short forms are the
3483            same. (This required a bit of code, as the output is generated
3484            automatically from a table. It wasn't just a text change.)
3485    
3486        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3487            option but starts with a hyphen. Could be a pattern or a path name
3488            starting with a hyphen, for instance.
3489    
3490        (h) "-" can be given as a file name to represent stdin.
3491    
3492        (i) When file names are being printed, "(standard input)" is used for
3493            the standard input, for compatibility with GNU grep. Previously
3494            "<stdin>" was used.
3495    
3496        (j) The option --label=xxx can be used to supply a name to be used for
3497            stdin when file names are being printed. There is no short form.
3498    
3499        (k) Re-factored the options decoding logic because we are going to add
3500            two more options that take data. Such options can now be given in four
3501            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3502    
3503        (l) Added the -A, -B, and -C options for requesting that lines of context
3504            around matches be printed.
3505    
3506        (m) Added the -L option to print the names of files that do not contain
3507            any matching lines, that is, the complement of -l.
3508    
3509        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3510            continue to scan other files.
3511    
3512        (o) The -s option was incorrectly implemented. For compatibility with other
3513            greps, it now suppresses the error message for a non-existent or non-
3514            accessible file (but not the return code). There is a new option called
3515            -q that suppresses the output of matching lines, which was what -s was
3516            previously doing.
3517    
3518        (p) Added --include and --exclude options to specify files for inclusion
3519            and exclusion when recursing.
3520    
3521    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3522        Hopefully, it now does.
3523    
3524    12. Missing cast in pcre_study().
3525    
3526    13. Added an "uninstall" target to the makefile.
3527    
3528    14. Replaced "extern" in the function prototypes in Makefile.in with
3529        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3530        world, but is set differently for Windows.
3531    
3532    15. Added a second compiling function called pcre_compile2(). The only
3533        difference is that it has an extra argument, which is a pointer to an
3534        integer error code. When there is a compile-time failure, this is set
3535        non-zero, in addition to the error test pointer being set to point to an
3536        error message. The new argument may be NULL if no error number is required
3537        (but then you may as well call pcre_compile(), which is now just a
3538        wrapper). This facility is provided because some applications need a
3539        numeric error indication, but it has also enabled me to tidy up the way
3540        compile-time errors are handled in the POSIX wrapper.
3541    
3542    16. Added VPATH=.libs to the makefile; this should help when building with one
3543        prefix path and installing with another. (Or so I'm told by someone who
3544        knows more about this stuff than I do.)
3545    
3546    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3547        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3548        match everything, including newlines. This is not POSIX-compatible, but
3549        somebody wanted the feature. From pcretest it can be activated by using
3550        both the P and the s flags.
3551    
3552    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3553    
3554    19. libpcre.pc was being incorrectly installed as executable.
3555    
3556    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3557        it now also looks for '\r' so that it will work unmodified on Windows.
3558    
3559    21. Added Google's contributed C++ wrapper to the distribution.
3560    
3561    22. Added some untidy missing memory free() calls in pcretest, to keep
3562        Electric Fence happy when testing.
3563    
3564    
3565    
3566    Version 5.0 13-Sep-04
3567    ---------------------
3568    
3569     1. Internal change: literal characters are no longer packed up into items
3570        containing multiple characters in a single byte-string. Each character
3571        is now matched using a separate opcode. However, there may be more than one
3572        byte in the character in UTF-8 mode.
3573    
3574     2. The pcre_callout_block structure has two new fields: pattern_position and
3575        next_item_length. These contain the offset in the pattern to the next match
3576        item, and its length, respectively.
3577    
3578     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3579        insertion of callouts before each pattern item. Added the /C option to
3580        pcretest to make use of this.
3581    
3582     4. On the advice of a Windows user, the lines
3583    
3584          #if defined(_WIN32) || defined(WIN32)
3585